CCP-C overview

  • 200 practice questions
  • 32 flashcards
  • 7 study guide topics
  • 39 key terms

What you'll study

  • Cloud Concepts and Architecture20%
  • Cloud Security, Risk, and Compliance20%
  • Deployment, Provisioning, and Operations18%
  • Networking and Connectivity12%
  • Data Management and Storage12%
  • Automation, DevOps, and Infrastructure as Code10%
  • Cost Management and Governance8%

CCP-C is a professional-level cloud certification aimed at practitioners who need to reason across architecture, security, operations, networking, data, automation, and cost - rather than specialize in a single vendor tool. It suits cloud engineers, systems administrators moving into cloud roles, solutions architects, and security or operations staff who share responsibility for cloud environments. Candidates typically arrive with hands-on exposure to at least one major cloud platform and a working grasp of networking, identity, and Linux or Windows administration.

The exam is multiple choice and scenario-driven. Questions tend to describe a business constraint - a compliance requirement, a latency target, a budget ceiling, a legacy dependency - and ask you to choose the design or operational response that best satisfies it. Content is organized into seven areas: cloud concepts and architecture; security, risk, and compliance; deployment, provisioning, and operations; networking and connectivity; data management and storage; automation, DevOps, and infrastructure as code; and cost management and governance. Expect the largest share of your study time to go to the first two.

Preparation works best in layers: build a mental model of each domain, then practise justifying trade-offs out loud, then drill the areas where your reasoning is slow. Reading alone is rarely enough - deploy something, break it, and fix it. Note that the code CCP-C is used by more than one credential family, so confirm the issuing organization and full certification title before you book anything; this page is independent study material and is not affiliated with or endorsed by any certifying body.

Common mistakes

Selecting an answer because the service name looks familiar, without first deciding whether the scenario describes IaaS, PaaS, or SaaS - for example, answering 'the customer must patch the guest operating system' on a scenario that actually describes a managed application platform.

Candidates memorise product catalogues and brand names instead of the service-model boundary, and the exam deliberately uses vendor-neutral wording that hides which layer is being managed by whom.

What to do instead

Before reading the options, write one line classifying the scenario against the five essential characteristics (on-demand self-service, broad network access, resource pooling, rapid elasticity, measured service) and the IaaS/PaaS/SaaS boundary. Then ask 'who manages the OS in this model?' and eliminate any option that contradicts it.

Treating high availability and disaster recovery as the same requirement - proposing a multi-region active-active design for a scenario that only asks to survive the loss of a single data centre within a stated availability target.

Both topics use the same vocabulary (redundancy, failover, resilience) and candidates pattern-match 'resilience' to 'multi-region' without checking the stated RTO, RPO, and availability figure.

What to do instead

Convert every scenario into explicit numbers first (RTO, RPO, availability %), then map the tiers: single-AZ = no HA, multi-AZ = HA within a region, multi-region = DR or extreme availability. Reject any option that exceeds the stated requirement, because over-engineering is scored as wrong when the stem says 'MOST cost-effective'.

Assuming the cloud provider secures everything above the physical layer - choosing options in which the provider patches the guest OS, hardens the application, or manages tenant identity and access configuration.

Marketing language such as 'fully managed' and 'secure by default' blurs the shared responsibility split, and candidates generalise from the provider-managed layer to the whole stack.

What to do instead

Draw the split per service model and check the option against it: for IaaS the provider owns facilities, hardware, and hypervisor while the customer owns OS, patching, and everything above; for PaaS the provider adds OS and runtime while the customer still owns application code, data, and identity/permission configuration; for SaaS the customer retains data, access management, and client-side configuration.

Conflating 'encryption is enabled' with 'the customer controls the key' - answering that provider-managed encryption satisfies a requirement for customer-controlled key rotation, key revocation, or the ability to make data unreadable on demand.

Encryption at rest, encryption in transit, and key custody are usually taught as one topic, so candidates tick the 'encrypted' option and stop reading the qualifying clause about key ownership.

What to do instead

Separate the three controls every time: (1) is the data encrypted at rest, (2) is it encrypted in transit with valid certificates, (3) who holds and can revoke the key (provider-managed, customer-managed/BYOK, or externally held HSM/HYOK). If the stem mentions revocation, rotation policy, or separation of duties, only a customer-controlled key option can be correct.

Mixing up release strategies - answering 'blue/green' for a stem that requires exposing a change to a small slice of live traffic while watching error rates, or answering 'canary' where the requirement is instant rollback by switching traffic back to the previous environment.

The strategies differ only in the traffic-handling detail, and candidates recall the labels without the trigger conditions that distinguish them.

What to do instead

Memorise each strategy by its trigger phrase: rolling = replace instances in batches, no extra capacity, mixed versions during rollout; blue/green = two full environments, instant cutover and instant rollback, double cost during the switch; canary = small percentage of traffic first, metrics-gated promotion. Match the stem's phrase ('minimise blast radius', 'instant rollback', 'no extra capacity') to the strategy before looking at the options.

Adding an explicit inbound rule for the return half of a connection because the candidate believes the cloud firewall is stateless - or, in the reverse error, assuming a subnet-level access control list automatically permits return traffic to ephemeral ports.

The two filtering layers look like the same feature in the console and the stateful/stateless distinction is a single word that candidates skip over when reading quickly.

What to do instead

Anchor the pair: security-group style filtering is stateful - allow the request and the response returns automatically; subnet-level ACLs are stateless - you must permit both directions, including the outbound request port and the inbound ephemeral range (commonly 1024–65535). Check the direction and the port range in the option before judging it.

Designing a multi-VPC topology with pairwise peering and assuming connectivity is transitive - for example, peering A-to-B and B-to-C and expecting A to reach C, or peering two networks whose IP ranges overlap.

Peering is described as 'private connectivity', so candidates mentally equate it with a routed network where every connected node reaches every other node.

What to do instead

Treat peering as a non-transitive one-to-one link. If the design has a hub, or more than two networks that all need to communicate, use a transit/hub construct (transit gateway, hub VPC with routing, or a shared network). Always verify that CIDR ranges do not overlap before selecting any peering option.

Accepting a read replica, a same-region snapshot, or a nightly export as the answer to a durability or disaster-recovery requirement - especially where the stated RPO is near zero or where the scenario describes accidental deletion or a malicious actor.

Replication and snapshots both look like 'copies of the data', so candidates stop at 'there is a copy' without checking where it lives, whether it is isolated from the primary account, whether it is immutable, and whether a restore has ever been tested.

What to do instead

Score each data-protection option on four axes: distance (same AZ, same region, cross-region), isolation (same account, cross-account, immutable/object-lock), lag (synchronous vs asynchronous, and the RPO it implies), and restore testability. Replication protects against infrastructure failure, not against deletion or corruption - only point-in-time recovery, versioning, or immutable copies do.

Treating infrastructure as code as a set of imperative deployment scripts - expecting a second run to 'just pick up the changes', hand-editing resources in the console afterwards, or planning changes without reviewing the diff first.

Candidates with scripting backgrounds map templates onto shell scripts, so idempotency, desired-state reconciliation, and state files are unfamiliar concepts that the options test directly.

What to do instead

Learn the four properties and check options against them: declarative (describe the end state, not the steps), idempotent (re-applying the same template changes nothing), stateful (a state file records what exists, and out-of-band console edits cause drift), and gated (a plan/diff is reviewed before apply). Any option that relies on manual console changes to a managed resource is drift, not a fix.

Choosing the cheapest published unit price for a steady-state 24/7 workload, and ignoring data transfer egress, cross-zone traffic, idle or oversized resources, storage tiering, and the staff time the option consumes.

The on-demand hourly rate is the most visible number, while egress charges and operational overhead are invisible in a quick comparison - and stems that say 'MOST cost-effective' invite exactly this shortcut.

What to do instead

Match workload pattern to pricing model first: spiky or short-lived = on-demand; predictable steady-state = commitment or savings plan; interruption-tolerant batch = spot/preemptible. Then build a mini total cost of ownership covering compute, storage tier chosen by access frequency, egress and cross-zone transfer, idle and orphaned resources, and the operational effort of running it. Re-read the stem's qualifier ('cost-effective' vs 'lowest operational overhead') before selecting.

Exam requirements can change. Confirm current details with the exam provider.

CCP-C practice

200 questions. Each one lists the choices, the correct answer, and an explanation. With JavaScript, this list opens one question at a time. The checker at the top of this page saves progress in this browser when it is available.

  1. Question 1Cloud Concepts and Architecture

    Which cloud characteristic lets a consumer start new server instances and storage without contacting the provider's staff?

    1. On-demand self-service Correct answer.
    2. Measured service
    3. Rapid elasticity
    4. Resource pooling

    Explanation

    On-demand self-service means the consumer unilaterally provisions capabilities such as compute and storage through a portal or API, with no human interaction on the provider side. The other characteristics describe how capacity scales, how resources are shared, and how usage is metered.

  2. Question 2Cloud Concepts and Architecture

    A development team wants to deploy only application code, leaving the operating system, runtime, and patches to the provider. Which service model fits best?

    1. Desktop as a service
    2. Infrastructure as a service
    3. Platform as a service Correct answer.
    4. Software as a service

    Explanation

    In platform as a service the provider manages servers, operating systems, and runtimes, while the consumer controls deployed applications and their configuration. Infrastructure as a service would leave the guest operating system in the consumer's hands.

  3. Question 3Cloud Concepts and Architecture

    Several agencies with shared regulatory requirements pool infrastructure that no single agency owns outright. Which deployment model does this describe?

    1. Private cloud
    2. Public cloud
    3. Hybrid cloud
    4. Community cloud Correct answer.

    Explanation

    A community cloud is provisioned for use by a specific group of organizations that share concerns such as compliance or mission, and it may be managed by the members or a third party. Public, private, and hybrid models differ in who the infrastructure serves and how it is owned.

  4. Question 4Cloud Concepts and Architecture

    A customer's monthly invoice reflects the number of API calls made and gigabytes stored. Which essential characteristic does this demonstrate?

    1. Resource pooling
    2. Rapid elasticity
    3. Measured service Correct answer.
    4. Broad network access

    Explanation

    Measured service means the provider meters resource usage and bills accordingly, giving both parties transparency into consumption. Resource pooling concerns sharing, broad network access concerns reachability, and rapid elasticity concerns scaling.

  5. Question 5Cloud Concepts and Architecture

    Users reach the same application from laptops, phones, and branch offices over the internet and dedicated links. Which characteristic is being demonstrated?

    1. Resource pooling
    2. Broad network access Correct answer.
    3. Measured service
    4. On-demand self-service

    Explanation

    Broad network access means capabilities are available over the network through standard mechanisms that promote use by heterogeneous client platforms. Self-service, pooling, and metering address provisioning, sharing, and billing rather than reachability.

  6. Question 6Cloud Concepts and Architecture

    A company stops buying servers upfront and instead pays a monthly cloud invoice for what it consumes. What does this shift in spending represent?

    1. Operating expenditure Correct answer.
    2. Capital expenditure
    3. Sunk cost
    4. Depreciation

    Explanation

    Cloud consumption is billed as a recurring operating expense rather than a large upfront capital investment in owned equipment. This shift converts capital expenditure into variable operating expenditure that tracks actual usage.

  7. Question 7Cloud Concepts and Architecture

    A team replaces a server with a larger one that has more vCPUs and memory. Which term describes this change?

    1. Auto-scaling
    2. Horizontal scaling
    3. Vertical scaling Correct answer.
    4. Elastic scaling

    Explanation

    Vertical scaling, or scaling up, increases the resources of a single instance. Horizontal scaling adds or removes instances, while auto-scaling is an automation policy that adjusts capacity in response to metrics.

  8. Question 8Cloud Concepts and Architecture

    Which storage type keeps data as discrete items with metadata, addressed over HTTP APIs, and scales without pre-provisioning capacity?

    1. Object storage Correct answer.
    2. File storage
    3. Block storage
    4. Tape archive

    Explanation

    Object storage stores each item as an object with data, metadata, and a unique key, accessed through HTTP APIs and scaling elastically across massive numbers of objects. Block and file storage expose volumes or shared hierarchies that behave like traditional disks or network shares.

  9. Question 9Cloud Concepts and Architecture

    Which statement best distinguishes scalability from elasticity?

    1. Scalability is capacity growth to meet load; elasticity automates growth and shrinkage. Correct answer.
    2. Scalability means shrinking capacity automatically, while elasticity means manually adding capacity during incidents.
    3. Scalability applies only to storage volumes, whereas elasticity applies only to compute instances.
    4. Both terms describe the same property, since cloud definitions treat scalability and elasticity as synonyms.

    Explanation

    Scalability is the capacity of a system to handle growing load, while elasticity adds automation so capacity expands and contracts with demand. Elastic systems are therefore scalable, but a manually scaled system is not necessarily elastic.

  10. Question 10Cloud Concepts and Architecture

    In a software-as-a-service deployment, which item typically remains the customer's responsibility?

    1. Patching the operating system that hosts the application
    2. Managing user identities and access Correct answer.
    3. Replacing failed physical disks in the provider's data center
    4. Maintaining the hypervisor layer

    Explanation

    In software as a service the provider operates the entire stack, so the customer's remaining duties are typically its own data, configuration, and the identities and access rights it grants to users. Infrastructure and platform layers stay with the provider.

  11. Question 11Cloud Concepts and Architecture

    An application must return to service within 4 hours of an outage and can lose at most 15 minutes of data. Which requirement pair is correct?

    1. RTO of 15 minutes and RPO of 15 minutes
    2. RTO of 4 hours and RPO of 15 minutes Correct answer.
    3. RTO of 15 minutes and RPO of 4 hours
    4. RTO of 4 hours and RPO of 4 hours

    Explanation

    Recovery time objective is the maximum acceptable time to restore service, so the 4-hour target is the RTO. Recovery point objective is the maximum acceptable data loss measured in time, so the 15-minute target is the RPO.

  12. Question 12Cloud Concepts and Architecture

    Which design most improves the availability of a stateless web tier inside a single region?

    1. Spreading the web instances across availability zones behind a load balancer Correct answer.
    2. Running every instance in one availability zone and taking hourly snapshots
    3. Replicating the database to another region and leaving the web tier unchanged
    4. Sizing a single instance large enough that it never reaches full utilization

    Explanation

    Availability zones are isolated failure domains within a region, so distributing instances across them keeps the tier serving traffic when one zone fails, with the load balancer routing around unhealthy targets. Snapshots address durability, not live availability.

  13. Question 13Cloud Concepts and Architecture

    Which party aggregates and integrates services from multiple providers and offers the combined capability to consumers?

    1. Cloud auditor
    2. Cloud consumer
    3. Cloud carrier
    4. Cloud broker Correct answer.

    Explanation

    In the reference actor model, a cloud broker manages service use and negotiates relationships between providers and consumers, often adding aggregation, integration, or customization. Carriers provide connectivity, auditors assess controls, and consumers use the services.

  14. Question 14Cloud Concepts and Architecture

    An application runs on-premises but automatically overflows to a public cloud when local capacity is exhausted, then returns when demand falls. What is this pattern called?

    1. Cloud bursting Correct answer.
    2. Hybrid brokerage
    3. Edge computing
    4. Multi-cloud federation

    Explanation

    Cloud bursting is a hybrid arrangement in which a private environment handles normal load and spills peak demand to a public cloud, usually for short periods. It depends on consistent images, networking, and data access so workloads can move between environments.

  15. Question 15Cloud Concepts and Architecture

    Which mechanism primarily allows a public provider to serve many customers from shared physical hosts while keeping their workloads separate?

    1. Encrypting stored data with customer-managed keys
    2. Replicating all customer data to every physical host
    3. Virtualization that isolates tenants on shared hosts Correct answer.
    4. Distributing incoming requests with a load balancer

    Explanation

    Resource pooling depends on virtualization, which carves physical hosts into isolated virtual instances so multiple tenants can share hardware without interfering with each other. Load balancing spreads traffic, and encryption protects data, but neither creates the tenant separation itself.

  16. Question 16Cloud Concepts and Architecture

    Which statement accurately describes how a container differs from a virtual machine?

    1. It always provides stronger isolation than a virtual machine
    2. It includes a full guest operating system and its own kernel
    3. It cannot be scheduled or orchestrated by a platform
    4. It shares the host kernel and bundles app dependencies Correct answer.

    Explanation

    A container packages an application with its libraries and dependencies but uses the host operating system kernel, which makes it lighter and faster to start than a virtual machine. A virtual machine runs a complete guest operating system with its own kernel.

  17. Question 17Cloud Concepts and Architecture

    Which component abstracts physical hosts so that virtual machines can be created, started, and isolated from one another?

    1. The API gateway
    2. The hypervisor Correct answer.
    3. The load balancer
    4. The object store

    Explanation

    The hypervisor runs on the physical host and allocates CPU, memory, and devices to isolated virtual machines. Gateways, load balancers, and object stores operate at higher layers and do not virtualize compute hardware.

  18. Question 18Cloud Concepts and Architecture

    Which workload is the best fit for a function-as-a-service platform?

    1. A legacy monolith that requires custom kernel modules
    2. An event-driven thumbnail generator invoked when images are uploaded Correct answer.
    3. A stateful database that must run continuously with tuned kernel parameters
    4. A 24/7 game server that keeps long-lived TCP connections open

    Explanation

    Function as a service suits short, event-triggered, stateless tasks that scale to zero when idle, such as processing an uploaded file. Long-running, stateful, or kernel-dependent workloads need persistent compute instead.

  19. Question 19Cloud Concepts and Architecture

    A website with a global audience wants to cut latency for static assets such as images and scripts. Which service should be added?

    1. A site-to-site VPN to each continent
    2. A content delivery network Correct answer.
    3. A block storage volume in the origin region
    4. A read replica of the relational database

    Explanation

    A content delivery network caches static content at points of presence near users, so requests are served from a nearby edge rather than the distant origin. Storage, VPNs, and database replicas do not cache static assets close to end users.

  20. Question 20Cloud Concepts and Architecture

    Under the shared responsibility model for infrastructure as a service, which task belongs to the customer?

    1. Replacing failed disks in the provider's storage array
    2. Maintaining the hypervisor and the host firmware
    3. Patching the guest OS and configuring security groups Correct answer.
    4. Securing the physical facility, including power and cooling

    Explanation

    In infrastructure as a service the provider secures facilities, hardware, and the hypervisor, while the customer owns everything from the guest operating system upward, including patching, network rules, and application security. The boundary sits at the virtualization layer.

  21. Question 21Cloud Concepts and Architecture

    A public web tier and a database sit in the same virtual network. Which design keeps the database unreachable from the internet while still allowing it to fetch software updates?

    1. Place the database in a public subnet but block the database port in a security group
    2. Assign the database a public IP address and rely on strong database authentication
    3. Put the database behind the public load balancer so only the load balancer can reach it
    4. Place the database in a private subnet with outbound access via a NAT gateway Correct answer.

    Explanation

    A private subnet has no route to an internet gateway, so inbound internet traffic cannot reach the database, while a NAT gateway allows outbound connections for updates. Address translation plus routing keeps the resource non-addressable from outside.

  22. Question 22Cloud Concepts and Architecture

    Each instance of a web tier stores session data on its local disk. What problem appears when the tier scales out behind a load balancer, and what is the standard remedy?

    1. The load balancer cannot register new instances; switch to DNS round robin
    2. Nothing changes; the load balancer replicates session data between instances
    3. Sessions are lost only at scale-in; increase the instance size to retain them
    4. Users may hit an instance without their session; externalize session state Correct answer.

    Explanation

    Local session state binds a user to one instance, so requests routed elsewhere lose the session and scale-in destroys it. Externalizing state to a shared store such as a cache or database makes the tier stateless and safe to scale horizontally.

  23. Question 23Cloud Concepts and Architecture

    Which scenario best illustrates portability rather than interoperability?

    1. Two clouds exchange data in real time through a shared API
    2. Both clouds use the same identity provider for single sign-on
    3. Both providers are used at once, connected by a shared private network
    4. A workload is exported and redeployed on another provider unchanged Correct answer.

    Explanation

    Portability is the ability to move a workload from one environment to another, so exporting and redeploying without code changes is the defining test. Interoperability concerns systems working together while remaining in place, such as shared APIs, identity, and networking.

  24. Question 24Cloud Concepts and Architecture

    Data gravity most directly shapes which architectural decision?

    1. Which programming language the front end should use
    2. Whether to require TLS 1.3 for all client connections
    3. How many availability zones a region should contain
    4. Where to place compute relative to large datasets Correct answer.

    Explanation

    Data gravity describes how large datasets attract the compute and services that use them, because moving data is slower and costlier than moving compute. Architects therefore co-locate processing with the bulk of the data and minimize cross-region transfer.

  25. Question 25Cloud Concepts and Architecture

    Which description matches a warm standby disaster recovery strategy?

    1. A scaled-down copy runs continuously and is scaled up during failover Correct answer.
    2. No cloud resources exist until a disaster is declared, then templates are deployed
    3. Only backups exist, and infrastructure is built from them after a disaster
    4. A full-size duplicate runs continuously and serves production traffic

    Explanation

    Warm standby keeps a functioning but reduced-capacity environment running, so failover requires scaling up rather than building from scratch, giving a moderate recovery time at moderate cost. It sits between pilot light and active-active in cost and speed.

  26. Question 26Cloud Concepts and Architecture

    A provider's services can be reached from a phone, a laptop, or an on-premises system over the internet using standard protocols. Which essential cloud characteristic does this describe?

    1. Broad network access Correct answer.
    2. Measured service
    3. Rapid elasticity
    4. Resource pooling

    Explanation

    Broad network access means capabilities are available over the network through standard mechanisms that promote use by heterogeneous client platforms. Measured service concerns metering and billing, resource pooling concerns shared physical hosts, and rapid elasticity concerns automatic capacity change.

  27. Question 27Cloud Concepts and Architecture

    Three hospitals with shared compliance requirements jointly fund a cloud environment that only they are allowed to use. Which deployment model is this?

    1. Community cloud Correct answer.
    2. Private cloud
    3. Public cloud
    4. Hybrid cloud

    Explanation

    A community cloud is provisioned for exclusive use by a specific community of consumers that share concerns such as compliance or mission. A private cloud serves a single organization, a public cloud is open to the general public, and a hybrid cloud combines two or more distinct models.

  28. Question 28Cloud Concepts and Architecture

    A development team wants to push application code to the cloud and have the provider handle the operating system, runtime, and patching. Which service model fits best?

    1. Software as a service
    2. Platform as a service Correct answer.
    3. Infrastructure as a service
    4. Desktop as a service

    Explanation

    Platform as a service supplies a managed runtime and middleware so developers deploy code while the provider manages servers, operating systems, and patching. Infrastructure as a service leaves operating system administration to the customer, software as a service delivers a finished application, and desktop as a service delivers virtual desktops.

  29. Question 29Cloud Concepts and Architecture

    Within a single cloud region, what does an availability zone represent?

    1. A cache node placed near end users to reduce latency
    2. A logical network segment that isolates tenants from one another
    3. A dedicated physical server reserved for a single customer
    4. Isolated data centers with separate power and cooling Correct answer.

    Explanation

    An availability zone is one or more physically separate data centers within a region, with independent power, cooling, and networking so a failure in one zone does not affect the others. It is not a per-customer server, a tenant network segment, or an edge cache.

  30. Question 30Cloud Concepts and Architecture

    A regulated company must keep personal data stored within its home country but wants fast response times for users abroad. Which design best satisfies both constraints?

    1. Replicate the entire database to every continent where users live
    2. Host all components in the single region with the most users
    3. Route each request to a random region through a global load balancer
    4. Keep the data in-country and add regional read caches near users Correct answer.

    Explanation

    Data residency requires the regulated data at rest to remain within the jurisdiction, so the system of record stays in-country while caches of permitted content near users absorb latency. Copying the regulated database to other continents breaks the constraint, a single region penalizes distant users, and random routing ignores both locality and residency.

  31. Question 31Cloud Concepts and Architecture

    A database server needs a low-latency volume that the operating system can format and treat like a locally attached disk. Which cloud storage type is appropriate?

    1. Block storage Correct answer.
    2. File storage
    3. Archive storage
    4. Object storage

    Explanation

    Block storage presents raw volumes that the guest operating system formats with its own file system, giving the consistent low-latency random I/O a database expects. Object storage is HTTP-addressed and cannot be mounted as a disk, archive tiers are for infrequent retrieval, and file storage adds a shared network file system layer.

  32. Question 32Cloud Concepts and Architecture

    A retail site sees traffic triple every weekday at 9 a.m. and fall again at 6 p.m., following a predictable pattern. Which scaling approach is the most efficient fit?

    1. Schedule capacity increases before 9 a.m. and decreases after 6 p.m. Correct answer.
    2. Scale up by moving to a larger instance size at each peak
    3. Provision permanently for the weekday peak all day and night
    4. Scale out only after CPU utilization has stayed above 80% for ten minutes

    Explanation

    Scheduled or predictive scaling adds capacity ahead of a known recurring demand pattern, so the workload never waits for a reactive threshold and does not pay for peak capacity around the clock. Reactive CPU rules lag behind a sharp, known spike, permanent peak provisioning wastes spend, and vertical scaling is disruptive and coarse.

  33. Question 33Cloud Concepts and Architecture

    A company needs a private, consistently low-latency link between its own data center and a public cloud provider, with predictable bandwidth for bulk replication. Which option fits best?

    1. A NAT gateway inside the cloud virtual network
    2. A dedicated private interconnect to the provider Correct answer.
    3. A site-to-site VPN over the public internet
    4. A content delivery network at the provider edge

    Explanation

    A dedicated interconnect bypasses the public internet and offers committed bandwidth with more predictable latency, which suits steady bulk replication. A VPN carries traffic over the internet where latency and throughput vary, a content delivery network caches content for end users, and a NAT gateway only translates outbound addresses for cloud resources.

  34. Question 34Cloud Concepts and Architecture

    Which architectural change most reduces the blast radius of a failure in one part of an application?

    1. Combining all functions into a single deployable unit
    2. Adding a second identical server behind a load balancer
    3. Splitting functions into independently deployed, asynchronously coupled services Correct answer.
    4. Running the existing application on a larger instance size

    Explanation

    Loose coupling means a failing service does not synchronously block its callers, so the effect of an outage stays contained to the affected function. A single deployable unit concentrates failures, a larger instance does not isolate faults, and a duplicate server protects only against the loss of that one whole server rather than one function.

  35. Question 35Cloud Concepts and Architecture

    Employees should sign in once with their existing corporate directory account and then reach several cloud providers without separate passwords. Which capability enables this?

    1. Federated identity with single sign-on Correct answer.
    2. Multi-factor authentication
    3. A shared administrator account
    4. Role-based access control

    Explanation

    Federation trusts the corporate identity provider to authenticate users and passes assertions to each cloud provider, giving single sign-on without duplicating credentials. Multi-factor authentication strengthens a single login but does not link providers, role-based access control governs permissions after authentication, and shared accounts undermine accountability.

  36. Question 36Cloud Concepts and Architecture

    An auto-scaling group must replace instances that freeze and stop responding. Which configuration achieves this?

    1. A scale-in policy triggered by low CPU utilization
    2. A longer instance startup timeout
    3. Load balancer health checks that mark failed instances for replacement Correct answer.
    4. A larger maximum group size

    Explanation

    Health checks probe each instance, and when a probe fails the orchestrator terminates and replaces it, which recovers frozen processes that a usage metric may never reveal. Scale-in removes healthy capacity, longer timeouts only delay start-up, and a larger maximum size merely raises the capacity ceiling.

  37. Question 37Cloud Concepts and Architecture

    A factory needs sensor analytics within milliseconds and cannot depend on a distant data center link. Which architecture addresses this?

    1. Move all processing to a centralized cloud region
    2. Batch the sensor data and upload it once per hour
    3. Process sensor data at an edge location close to the equipment Correct answer.
    4. Increase bandwidth on the wide-area link to the cloud

    Explanation

    Edge computing places compute near the data source, removing round-trip latency to a distant region and allowing the workload to continue when the wide-area link fails. Central processing adds latency, extra bandwidth does not remove propagation delay, and batching defers results rather than accelerating them.

  38. Question 38Cloud Concepts and Architecture

    A disaster recovery plan keeps only a minimal core of the production environment always running in a second region, with the remaining components switched on during a failover. Which strategy is this?

    1. Warm standby
    2. Pilot light Correct answer.
    3. Multi-site active-active
    4. Backup and restore

    Explanation

    A pilot light replicates only the core, such as data and minimal configuration, and starts the rest of the stack at failover, which costs less than warm standby but recovers more slowly. Backup and restore starts from nothing, warm standby keeps a scaled-down but fully functional stack running, and active-active serves traffic from both regions continuously.

  39. Question 39Cloud Concepts and Architecture

    In a cloud ecosystem, which role provides the network connectivity that transports services between a provider and its consumers?

    1. Cloud consumer
    2. Cloud auditor
    3. Cloud carrier Correct answer.
    4. Cloud broker

    Explanation

    The cloud carrier is the intermediary that provides connectivity and transport of cloud services between providers and consumers. A broker aggregates and integrates services, an auditor assesses controls and compliance, and the consumer uses the services.

  40. Question 40Cloud Concepts and Architecture

    Which comparison best explains why a like-for-like cloud bill can still be lower than an on-premises total cost of ownership?

    1. Cloud providers do not charge for data transfer between regions
    2. Cloud resources never require operational staff
    3. On-premises estimates often omit staff, power, facilities and refresh Correct answer.
    4. On-premises hardware is always more expensive per CPU hour

    Explanation

    A fair total cost of ownership comparison includes power, cooling, floor space, administrative labor, and periodic hardware refresh, which are frequently left out of on-premises figures and are absorbed into cloud prices. Transfer fees do exist, cloud workloads still need operations staff, and per-hour component prices alone do not capture total cost.

  41. Question 41Cloud Security, Risk, and Compliance

    Under the cloud shared responsibility model, which activity always remains the customer's responsibility regardless of whether the service is IaaS, PaaS, or SaaS?

    1. Securing the physical facilities that host the servers
    2. Maintaining the underlying network fabric
    3. Classifying data and managing who can access it Correct answer.
    4. Patching the hypervisor and host operating systems

    Explanation

    Providers secure the facilities, hardware, and, depending on the service model, the hypervisor, runtime, and application. Data, identity, and access decisions stay with the customer in every model. Classifying data and governing who can reach it is therefore never delegated to the provider.

  42. Question 42Cloud Security, Risk, and Compliance

    Which program provides a standardized security assessment and authorization that U.S. federal agencies rely on when adopting cloud services?

    1. PCI DSS
    2. FedRAMP Correct answer.
    3. SOC 2 Type II
    4. ISO/IEC 27001

    Explanation

    FedRAMP is a government-wide program that standardizes assessment, authorization, and continuous monitoring for cloud products and services used by federal agencies. SOC 2 is a private attestation report, PCI DSS focuses on cardholder data, and ISO/IEC 27001 is an international management-system standard.

  43. Question 43Cloud Security, Risk, and Compliance

    A regulator requires that customer records never be stored outside the country of origin. Which concept does this requirement describe?

    1. Data masking
    2. Data residency Correct answer.
    3. Data retention
    4. Data minimization

    Explanation

    Data residency is the requirement that data be stored within a specified geographic or jurisdictional boundary. Minimization limits what is collected, retention limits how long it is kept, and masking hides values while the record remains in place.

  44. Question 44Cloud Security, Risk, and Compliance

    Which practice best implements the principle of least privilege for cloud identities?

    1. Granting administrator rights to all engineers
    2. Sharing a single service account across teams
    3. Storing long-lived access keys in source code
    4. Granting only the permissions a role needs Correct answer.

    Explanation

    Least privilege means each identity receives the minimum permissions required for its function, with access reviewed and removed when no longer needed. Standing administrator rights, shared accounts, and embedded long-lived keys all exceed that minimum and expand the blast radius of a compromise.

  45. Question 45Cloud Security, Risk, and Compliance

    What is the primary purpose of classifying data held in cloud services?

    1. To reduce the storage cost of the data
    2. To increase the number of backups taken
    3. To eliminate the need for encryption
    4. To apply controls proportionate to its sensitivity Correct answer.

    Explanation

    Classification tells an organization how sensitive an asset is so that controls such as encryption, access restrictions, and retention can be matched to that sensitivity. Cost, backup frequency, and encryption decisions may follow from classification but are not its purpose.

  46. Question 46Cloud Security, Risk, and Compliance

    Which statement accurately describes a SOC 2 report?

    1. It guarantees a vendor will never suffer a breach
    2. It is a technical vulnerability scan of the network
    3. It is a government certification that remains valid for three years
    4. It attests to controls against the trust services criteria Correct answer.

    Explanation

    A SOC 2 report is an independent attestation on a service organization's controls relevant to one or more trust services criteria, such as security, availability, and confidentiality, covering a defined period. It is not a government certification, a vulnerability scan, or a guarantee against future incidents.

  47. Question 47Cloud Security, Risk, and Compliance

    An organization wants to retain control of the encryption keys for data it stores with a cloud provider. Which approach fits that goal?

    1. Disabling encryption so keys are not needed
    2. Relying only on provider-managed keys without an agreement
    3. Storing the keys in the same bucket as the encrypted data
    4. Using a customer-managed key held in an external key vault Correct answer.

    Explanation

    A customer-managed key held outside the provider's control lets the organization govern key lifecycle and revoke access by disabling the key. Keeping keys beside the ciphertext collapses the separation that makes encryption meaningful, and provider-managed default keys leave control with the provider.

  48. Question 48Cloud Security, Risk, and Compliance

    Which control protects data moving between a user's browser and a cloud-hosted application?

    1. Object versioning on the storage bucket
    2. Full-disk encryption on the application server
    3. TLS on every client connection Correct answer.
    4. Column-level masking in the database

    Explanation

    Transport layer security encrypts data in transit and prevents interception or tampering between client and server. Disk encryption, versioning, and column masking all act on stored data and do nothing for the network path.

  49. Question 49Cloud Security, Risk, and Compliance

    Which cloud service model places the greatest security responsibility on the customer?

    1. Functions as a Service
    2. Infrastructure as a Service Correct answer.
    3. Platform as a Service
    4. Software as a Service

    Explanation

    In IaaS the customer manages the guest operating system, middleware, runtime, applications, and data, while the provider handles only the physical and virtualization layers. In PaaS, SaaS, and serverless models the provider absorbs progressively more of the stack.

  50. Question 50Cloud Security, Risk, and Compliance

    A company processes personal data of individuals located in the European Union. Which obligation arises under the GDPR?

    1. Retain all personal data indefinitely so audits can be performed
    2. Obtain consent only after the data has been shared
    3. Notify the supervisory authority of qualifying breaches within 72 hours Correct answer.
    4. Transfer data to any country without additional safeguards

    Explanation

    The GDPR requires notification of qualifying personal data breaches to the competent supervisory authority without undue delay and, where feasible, within 72 hours of awareness. It also imposes storage limitation and restricts transfers to third countries without an appropriate safeguard.

  51. Question 51Cloud Security, Risk, and Compliance

    An alert indicates that a cloud access key may be compromised. What is the most appropriate immediate containment step?

    1. Defer action until the next scheduled security review
    2. Publish a public statement before investigating
    3. Delete the account's log trail to stop further tampering
    4. Disable the key and terminate its active sessions Correct answer.

    Explanation

    Containment must happen first: disabling or rotating the credential and invalidating active sessions stops the attacker from continuing to use it. Investigation, evidence preservation, and notification follow containment.

  52. Question 52Cloud Security, Risk, and Compliance

    Before onboarding a SaaS vendor that will hold sensitive customer data, what should the security team do first?

    1. Migrate the data so testing can begin sooner
    2. Assess the vendor's security controls and attestations Correct answer.
    3. Assume the vendor is secure because it is a large company
    4. Sign the contract before any security evidence is reviewed

    Explanation

    Third-party risk management requires due diligence before data is shared, including review of the vendor's controls, independent attestations, and incident history. Organization size and contractual urgency are not evidence of security maturity.

  53. Question 53Cloud Security, Risk, and Compliance

    Why should cloud audit logs be shipped to a separate, append-only account?

    1. To remove the need for any log retention policy
    2. To keep evidence out of reach of compromised workloads Correct answer.
    3. To improve the response time of the application
    4. To reduce the monthly spend on the logging tier

    Explanation

    An attacker who compromises a workload often has enough permission to delete or disable logging in that same account. Shipping logs to a separate account with write-only access from the source preserves evidence and satisfies the separation-of-duties expectation behind most audit controls.

  54. Question 54Cloud Security, Risk, and Compliance

    A business states that after an outage it can tolerate losing at most fifteen minutes of transactions. Which metric does this define?

    1. Mean time to repair
    2. Recovery point objective Correct answer.
    3. Recovery time objective
    4. Service level agreement

    Explanation

    The recovery point objective defines the maximum acceptable amount of data loss, measured in time, and therefore drives backup and replication frequency. The recovery time objective defines how quickly service must be restored, which is a separate commitment.

  55. Question 55Cloud Security, Risk, and Compliance

    Which control is most appropriate for detecting account numbers being emailed outside the organization?

    1. A web application firewall
    2. Data loss prevention scanning of outbound email Correct answer.
    3. Full-disk encryption for all corporate laptops and desktops
    4. Multi-factor authentication

    Explanation

    Data loss prevention inspects outbound content against patterns and policy to detect or block sensitive data leaving the organization. A web application firewall filters HTTP traffic to applications, MFA strengthens authentication, and disk encryption protects stored data on devices.

  56. Question 56Cloud Security, Risk, and Compliance

    At which point in the development lifecycle is threat modeling most cost-effective for a cloud service?

    1. During design, before code is written Correct answer.
    2. During the annual financial audit cycle
    3. Only after a penetration test finds a critical flaw
    4. After the service has run in production for a year

    Explanation

    Threat modeling identifies trust boundaries, entry points, and likely attack paths while architecture is still malleable, when remediation costs the least. Once code is written or the service is live, design changes become far more expensive.

  57. Question 57Cloud Security, Risk, and Compliance

    Which type of data does PCI DSS primarily exist to protect?

    1. Employee health records
    2. Source code repositories
    3. Payment card account data Correct answer.
    4. Government classified documents

    Explanation

    PCI DSS sets requirements for organizations that store, process, or transmit cardholder data and sensitive authentication data. Health records fall under HIPAA, classified documents under government handling rules, and source code under intellectual property controls.

  58. Question 58Cloud Security, Risk, and Compliance

    What is the Cloud Controls Matrix published by the Cloud Security Alliance?

    1. A set of cloud security controls mapped to frameworks Correct answer.
    2. A public price list comparing the major cloud providers
    3. A standard contract template for buying cloud services
    4. A commercial vulnerability scanner for container images

    Explanation

    The Cloud Controls Matrix is a control framework specific to cloud computing that is cross-referenced to widely used standards, which makes it useful for gap analysis and vendor questionnaires. It is a control catalog, not a pricing list, scanning tool, or contract.

  59. Question 59Cloud Security, Risk, and Compliance

    In a qualitative cloud risk assessment, what is the usual purpose of a risk matrix?

    1. To enumerate every asset in the environment
    2. To rank risks by combining likelihood and impact Correct answer.
    3. To calculate the exact annual loss expectancy in currency
    4. To record the patch level of each server

    Explanation

    A risk matrix plots likelihood against impact so risks can be compared and prioritized for treatment. It produces relative rankings, not precise monetary figures, and it sits above asset inventories and configuration data rather than replacing them.

  60. Question 60Cloud Security, Risk, and Compliance

    Why is strong tenant isolation a critical property of a public cloud environment?

    1. It allows tenants to share the same encryption keys safely
    2. It removes the need for access logging
    3. It prevents one tenant from accessing another's data Correct answer.
    4. It lowers the cost of the provider's storage systems

    Explanation

    Public cloud resources share physical infrastructure, so logical isolation of compute, network, and storage is what keeps one customer's activity from reaching another's data. Shared keys and reduced logging would undermine rather than support that boundary.

  61. Question 61Cloud Security, Risk, and Compliance

    A firm requires that the cloud provider be unable to decrypt its stored data even under legal compulsion. Which design best meets that requirement?

    1. Enabling provider-managed encryption at rest for every volume
    2. Relying on TLS for all API calls to the storage service
    3. Using provider-managed keys with automatic annual rotation
    4. Encrypting with keys held only in the customer's on-premises HSM Correct answer.

    Explanation

    If key material never leaves the customer's own hardware security module, the provider holds only ciphertext and cannot decrypt it. Provider-managed keys, even with rotation or default encryption enabled, remain accessible to the provider, and TLS protects the channel rather than the stored object.

  62. Question 62Cloud Security, Risk, and Compliance

    During a cloud security incident, which action best preserves evidence for later forensic analysis?

    1. Repeatedly rebooting the instance so the fault can be reproduced
    2. Reimaging the host with a clean operating system image
    3. Terminating the compromised instance as quickly as possible
    4. Snapshotting volumes and copying logs to a restricted account Correct answer.

    Explanation

    Forensic soundness depends on capturing volatile and persistent state before it changes: disk snapshots, memory where possible, and immutable copies of logs in a controlled account. Rebooting, terminating, or reimaging destroys the artifacts an investigation needs.

  63. Question 63Cloud Security, Risk, and Compliance

    A data subject asks an organization to erase personal data that exists in primary storage, backups, and logs. What is the most defensible approach?

    1. Wait until the next annual audit to review the request
    2. Erase or anonymize every reachable copy under a documented process Correct answer.
    3. Erase only the primary database record and nothing else
    4. Inform the customer that cloud erasure is technically impossible

    Explanation

    Erasure obligations extend to personal data across systems, including derived copies, with backups handled through a documented cycle so data does not silently reappear on restore. A repeatable, documented process is what demonstrates compliance and accountability.

  64. Question 64Cloud Security, Risk, and Compliance

    What does crypto-shredding accomplish when data must be securely deleted from cloud storage?

    1. It arranges physical destruction of the provider's storage hardware
    2. It destroys the keys so the remaining ciphertext is unreadable Correct answer.
    3. It overwrites every disk block with random data across all replicas
    4. It deletes object metadata so the console no longer lists the object

    Explanation

    With envelope encryption, deleting the data encryption key renders all copies of the ciphertext permanently unreadable without needing to reach every replica or backup. Overwriting or physically destroying media is impractical in multi-tenant cloud storage, and deleting metadata leaves the bytes intact.

  65. Question 65Cloud Security, Risk, and Compliance

    A cloud contract states the provider will notify the customer of a breach within a reasonable time. Why is this wording a compliance concern?

    1. It transfers all liability for the breach to the customer
    2. It requires the customer to encrypt the data twice
    3. It prevents the customer from deploying into multiple regions
    4. It may not satisfy a regulatory breach notification deadline Correct answer.

    Explanation

    Many regimes require the data controller to notify regulators or affected individuals within a fixed window, so a vague provider commitment can make that deadline impossible to meet. Contracts should specify a concrete notification period and the content of the notice.

  66. Question 66Cloud Security, Risk, and Compliance

    A company decides to buy a cyber insurance policy covering losses from a cloud outage instead of funding a second active region. Which risk treatment option has it chosen?

    1. Risk transfer, by shifting the financial consequence to an insurer Correct answer.
    2. Risk avoidance, by removing the reliance on a single region
    3. Risk mitigation, by adding controls that lower the likelihood of the event
    4. Risk acceptance, by acknowledging the exposure without further action

    Explanation

    Risk treatment options are commonly grouped as mitigate, transfer, avoid, and accept. Buying insurance does not change the likelihood or impact of the outage; it shifts the financial consequence to a third party, which is the definition of transfer.

  67. Question 67Cloud Security, Risk, and Compliance

    Why does an organization assign sensitivity tiers to data held in cloud storage?

    1. To satisfy provider licensing terms for hosted workloads
    2. To measure the throughput of the encryption appliances in use
    3. To lower the storage cost of data that is rarely accessed
    4. To determine which handling and protection controls each data set requires Correct answer.

    Explanation

    Classification maps data to a sensitivity level so that proportionate controls such as encryption, access restrictions, and retention rules can be selected. Without a classification scheme, control strength is applied uniformly or arbitrarily rather than according to value and regulatory exposure.

  68. Question 68Cloud Security, Risk, and Compliance

    After selected controls have been implemented against a cloud risk, what does the residual risk represent?

    1. The risk that still remains once the chosen controls are in place Correct answer.
    2. The maximum loss an insurer will reimburse under a policy
    3. The portion of the risk that the cloud provider contractually owns
    4. The risk that exists before any controls have been selected

    Explanation

    Residual risk is the exposure left after controls have been applied, and it is what management formally accepts or escalates. It is compared against the organization's risk tolerance to decide whether further treatment is needed.

  69. Question 69Cloud Security, Risk, and Compliance

    What is the primary role of a cloud access security broker (CASB)?

    1. Encrypting block storage volumes attached to virtual machines
    2. Enforcing policy and giving visibility over data moving to and within cloud services Correct answer.
    3. Approving which cloud regions may host regulated workloads
    4. Replacing the identity provider for federated sign-in to cloud services

    Explanation

    A CASB sits between users and cloud services and applies policy such as data loss prevention, access control, and shadow-IT discovery. It observes and governs cloud traffic rather than performing infrastructure-level functions.

  70. Question 70Cloud Security, Risk, and Compliance

    Which subject does ISO/IEC 27018 specifically address?

    1. Security controls for cloud services and the provider-customer relationship
    2. Protection of personally identifiable information in public clouds acting as processors Correct answer.
    3. Assessment of cloud providers against the Cloud Controls Matrix
    4. Generic requirements for an information security management system

    Explanation

    ISO/IEC 27018 is the code of practice for protecting personally identifiable information in public clouds where the provider acts as a processor. It extends the generic cloud controls with privacy-specific obligations such as purpose limitation and disclosure transparency.

  71. Question 71Cloud Security, Risk, and Compliance

    A customer wants independent third-party assurance that a cloud provider's controls align with the Cloud Controls Matrix. Which CSA STAR level should it request?

    1. Level 2, an independent third-party assessment or certification Correct answer.
    2. No STAR level applies, because STAR covers only privacy questionnaires
    3. A STAR level that relies only on the provider's internal audit team
    4. Level 1, a self-assessment published by the provider

    Explanation

    STAR is tiered by assurance strength: Level 1 is a provider self-assessment, Level 2 adds independent third-party assessment or certification, and Level 3 adds continuous monitoring. A request for third-party assurance therefore maps to Level 2.

  72. Question 72Cloud Security, Risk, and Compliance

    A SaaS provider plans to add a new subprocessor that will handle customer personal data. Under typical GDPR-aligned contracting, what right does the customer have?

    1. To contract directly with the subprocessor and bypass the provider
    2. To select the subprocessor's staff and review their employment records
    3. To require all data to be hosted in the subprocessor's home country
    4. To be informed in advance of the addition and to object to or refuse it Correct answer.

    Explanation

    Where a processor engages another processor, the controller must authorize the arrangement and be given the chance to object to changes. Contracts therefore require prior notice of new subprocessors plus an objection or approval mechanism.

  73. Question 73Cloud Security, Risk, and Compliance

    A security team intends to run a penetration test against workloads it hosts with a public cloud provider. What should it do first?

    1. Ask the provider to run the assessment on the customer's behalf
    2. Test the shared hypervisor layer to validate tenant isolation
    3. Review and follow the provider's testing policy, including any notification or approval requirements Correct answer.
    4. Start testing at once, because customer workloads are always in scope

    Explanation

    Cloud providers publish acceptable-use and penetration-testing policies that define permitted targets, techniques, and notice requirements. Complying with those terms first avoids breaching the service agreement and avoids testing infrastructure the customer does not own.

  74. Question 74Cloud Security, Risk, and Compliance

    A cloud provider holds an ISO/IEC 27001 certificate covering its data centers. What is the sound conclusion for a customer reviewing it?

    1. It supports the customer's assessment but does not replace the customer's own controls Correct answer.
    2. The customer inherits full compliance with its own regulatory obligations
    3. The certificate covers the customer's applications deployed on that infrastructure
    4. The customer can stop assessing the provider once a copy is on file

    Explanation

    Provider certifications are evidence about the provider's control environment within a defined scope, and they support the customer's due diligence. Responsibility for the customer's own configurations, identity management, and regulatory obligations remains with the customer.

  75. Question 75Cloud Security, Risk, and Compliance

    A regulator insists that records about its nationals remain subject to national law even when processed in another country. Which concept does this describe?

    1. Data sovereignty, which concerns which jurisdiction's law governs the data Correct answer.
    2. Data residency, which concerns the physical location of storage
    3. Data remanence, which concerns copies left behind on storage media
    4. Data minimization, which concerns collecting only what is necessary

    Explanation

    Sovereignty is about legal control, meaning the laws of a state apply to the data regardless of where it physically sits. Residency is the narrower logistical question of which country's facilities store it.

  76. Question 76Cloud Security, Risk, and Compliance

    Why might a customer still negotiate a right-to-audit clause when the provider already publishes independent audit reports?

    1. Because a right-to-audit removes the need for periodic reassessment
    2. Because published reports transfer the provider's liability to the customer under contract
    3. Because the clause provides recourse beyond a report's fixed scope and period Correct answer.
    4. Because audit reports are binding certifications of the customer's own controls

    Explanation

    Published reports describe controls as of a stated period and scope, which may not match every customer's concerns or current conditions. A contractual audit right gives the customer a route to obtain additional assurance, such as questionnaires or onsite review, when questions fall outside the report.

  77. Question 77Cloud Security, Risk, and Compliance

    What is the primary purpose of a cloud risk register?

    1. To track each identified risk, its owner, rating, and treatment status Correct answer.
    2. To list every cloud asset together with its configuration baseline
    3. To document the provider's service level commitments and credits
    4. To record the findings of the most recent penetration test

    Explanation

    A risk register is the living record that lets management see what risks exist, who owns them, how they are rated, and what treatment decision was made. It supports prioritization, accountability, and reporting to governance bodies over time.

  78. Question 78Cloud Security, Risk, and Compliance

    A customer reviews a provider's audit report several months after the report period closed and receives a bridging letter with it. What does the letter provide?

    1. It retroactively extends the audited period to cover the missing months
    2. It substitutes for the full report for customers in another regulatory region
    3. It certifies the customer's own complementary controls for the same period
    4. It states whether the reported controls changed during the gap period Correct answer.

    Explanation

    A bridging letter covers the interval between the end of the report period and the date of the customer's review. It tells the reader whether controls described in the report have changed materially and whether the report can still be relied upon for that gap.

  79. Question 79Cloud Security, Risk, and Compliance

    An organization relies on standard contractual clauses to transfer EU personal data to a cloud region in a third country. What must accompany that reliance?

    1. Deletion of all encryption keys held by the exporting organization
    2. An assessment of the destination's legal regime and any supplementary safeguards Correct answer.
    3. A fresh Cloud Controls Matrix certification of the provider for that region
    4. Nothing further, because the clauses take effect automatically once signed

    Explanation

    Transfer mechanisms must be supported by an evaluation of whether the destination's laws and practices undermine the protections the clauses promise. Where they do, supplementary technical and organizational measures must be added before the transfer proceeds.

  80. Question 80Cloud Security, Risk, and Compliance

    A company is negotiating a cloud agreement and worries that the provider could become insolvent. Which provision best protects its ability to leave?

    1. An unlimited liability cap for security breaches
    2. A service credit schedule tied to monthly availability targets
    3. Transition assistance with export in a standard format and a wind-down period Correct answer.
    4. A most-favored-customer pricing commitment

    Explanation

    Exit risk is addressed through provisions that guarantee retrievability and continued access during a transition, including usable export formats and a defined wind-down window. Financial remedies such as credits or liability caps do not restore access to data once the service ends.

  81. Question 81Deployment, Provisioning, and Operations

    Which characteristic best describes infrastructure as code (IaC)?

    1. Each environment is configured by the engineer who happens to own it
    2. Resources are defined in version-controlled template files Correct answer.
    3. Servers are rebuilt manually from written runbooks whenever a change is needed
    4. Capacity is purchased only after a procurement ticket has been approved

    Explanation

    IaC treats infrastructure definitions as source code stored in a repository, so every change is reviewable, repeatable, and reversible. The template file is the record of intent, and tooling applies it to the environment.

  82. Question 82Deployment, Provisioning, and Operations

    A recovery point objective (RPO) specifies which of the following?

    1. The number of geographic regions a workload must span
    2. The maximum time a service may be unavailable after an outage
    3. The maximum acceptable amount of data loss measured in time Correct answer.
    4. The interval at which backups are encrypted

    Explanation

    The RPO is expressed as a period of time and states how much data the business can afford to lose, which in turn drives backup and replication frequency. It is a data-loss bound, not an availability bound.

  83. Question 83Deployment, Provisioning, and Operations

    In a blue-green deployment, what does the team switch at cutover?

    1. Traffic routing from the old environment to the new environment Correct answer.
    2. The source control repository that builds the release
    3. The container orchestration platform that hosts the workload
    4. The database engine version used by the application

    Explanation

    Blue-green keeps two complete environments and flips the routing layer so users move from the old stack to the new one in a single controlled step. Rollback is fast because the previous environment remains intact.

  84. Question 84Deployment, Provisioning, and Operations

    Adding more instances of an application to handle increased demand is an example of which scaling approach?

    1. Vertical scaling
    2. Elastic caching
    3. Horizontal scaling Correct answer.
    4. Manual capacity reservation

    Explanation

    Horizontal scaling, also called scaling out, adds more nodes to the pool so work is distributed across them. Vertical scaling instead increases the size of a single node.

  85. Question 85Deployment, Provisioning, and Operations

    A load balancer is configured with health checks against each backend instance. What is the primary purpose of these checks?

    1. To encrypt traffic between the client and the load balancer
    2. To stop routing traffic to instances that are failing Correct answer.
    3. To compress responses before they leave the network
    4. To measure how much each instance costs to operate

    Explanation

    Health checks continuously probe backends so the load balancer can remove unhealthy targets from rotation and return them when they recover. This keeps requests away from instances that cannot serve them.

  86. Question 86Deployment, Provisioning, and Operations

    Which tagging practice most directly supports cloud cost allocation?

    1. Tagging only resources that are created through the console
    2. Tagging resources after they have been deleted
    3. Using a different tag key spelling for each team
    4. Applying consistent tags such as cost center, owner, and environment Correct answer.

    Explanation

    Cost allocation depends on a consistent tag taxonomy applied to every resource, so billing data can be grouped by owner, environment, or cost center. Inconsistent or partial tagging leaves spend unattributable.

  87. Question 87Deployment, Provisioning, and Operations

    An operation that produces the same end state no matter how many times it is executed is described as what?

    1. Elastic
    2. Stateless
    3. Idempotent Correct answer.
    4. Eventually consistent

    Explanation

    Idempotency means repeated execution converges on the same result without side effects such as duplicate resources. Provisioning and configuration tools rely on it so retries are safe.

  88. Question 88Deployment, Provisioning, and Operations

    Which mechanism is commonly used to run initial configuration tasks on a newly provisioned virtual machine at first boot?

    1. A support ticket sent to the provider
    2. A cron job scheduled a week in advance
    3. Instance user data processed by a bootstrap agent Correct answer.
    4. A manual SSH session opened by an operator

    Explanation

    User data or metadata supplied at launch is read by a bootstrap agent during first boot and used to install packages, join clusters, or fetch configuration. This makes initial setup automated and repeatable.

  89. Question 89Deployment, Provisioning, and Operations

    A team routes 5% of production traffic to a new release and compares error rates against the baseline before widening exposure. Which deployment strategy is this?

    1. Shadow deployment
    2. Blue-green cutover
    3. Canary release Correct answer.
    4. Recreate deployment

    Explanation

    A canary release exposes a small slice of live traffic to the new version and expands that slice only if metrics stay healthy. Blue-green instead moves all traffic at once, and shadow deployments never serve user responses from the new version.

  90. Question 90Deployment, Provisioning, and Operations

    Why do teams that practice immutable infrastructure replace failed instances instead of patching them in place?

    1. Because replaced instances are guaranteed to match the tested image Correct answer.
    2. Because replacement always costs less than patching
    3. Because instances cannot be patched in any cloud provider
    4. Because patching requires downtime for every workload

    Explanation

    With immutable infrastructure the server image is the unit of change, so a replacement is built from the same tested artifact that passed the pipeline. This removes configuration drift and makes rollback a simple matter of launching the prior image.

  91. Question 91Deployment, Provisioning, and Operations

    After an emergency manual change in the console, an IaC apply proposes to revert that change. What does this indicate?

    1. The provider plugin is out of date
    2. The state file has been deleted and must be recreated
    3. The template contains a syntax error
    4. The live environment has drifted from the declared configuration Correct answer.

    Explanation

    Configuration drift occurs when the live environment no longer matches the declared template, and the tool's plan reflects the difference it intends to reconcile. Reverting the manual change restores the declared state.

  92. Question 92Deployment, Provisioning, and Operations

    Where should an application running in the cloud retrieve database credentials at runtime?

    1. In a plaintext configuration file that is committed to the source repository
    2. From a managed secrets store accessed with a scoped identity Correct answer.
    3. In the container image build layer
    4. In an environment variable printed in the logs

    Explanation

    Secrets should live in a dedicated store that enforces access control, rotation, and audit, and the workload should authenticate with an identity scoped to only that secret. This keeps credentials out of code, images, and log streams.

  93. Question 93Deployment, Provisioning, and Operations

    A service has a recovery time objective of two hours. What must the recovery plan guarantee?

    1. No more than two hours of data is lost
    2. Service is restored within two hours of the outage Correct answer.
    3. Backups are taken every two hours
    4. The outage is announced within two hours

    Explanation

    The recovery time objective bounds how long the service may remain unavailable, so the recovery procedure must complete inside that window. Data loss is bounded separately by the recovery point objective.

  94. Question 94Deployment, Provisioning, and Operations

    In a rolling deployment, why is a maximum-unavailable setting used?

    1. To limit how much capacity is removed at once during the rollout Correct answer.
    2. To cap the total number of releases per quarter
    3. To restrict which engineers may approve the release
    4. To bound the storage consumed by container images

    Explanation

    A rolling deployment replaces instances in batches, and the maximum-unavailable value controls how many may be offline simultaneously so the service retains enough capacity. It balances rollout speed against availability.

  95. Question 95Deployment, Provisioning, and Operations

    An application must survive the loss of a single data center. Which architecture satisfies this with the least change?

    1. A standby copy of the database in the same rack as the primary
    2. Instances spread across availability zones behind a load balancer Correct answer.
    3. Backups written nightly and restored into the same data center after an incident
    4. A single instance with a large instance type

    Explanation

    Availability zones are isolated failure domains, so spreading instances across zones means the loss of one site does not take down the service. The load balancer then routes only to the surviving healthy targets.

  96. Question 96Deployment, Provisioning, and Operations

    Which task belongs to configuration management rather than to initial resource provisioning?

    1. Enforcing installed package versions on running hosts Correct answer.
    2. Allocating a public IP address
    3. Attaching a block storage volume
    4. Creating the virtual network and subnets

    Explanation

    Configuration management keeps running systems in a desired software state, including package versions, services, and file content. Provisioning creates the infrastructure itself, such as networks, addresses, and storage.

  97. Question 97Deployment, Provisioning, and Operations

    An operations team needs to know when average request latency exceeds a threshold for five minutes. Which capability provides this?

    1. An audit log of API calls
    2. A static dashboard refreshed on demand
    3. A metric alarm with a defined evaluation period Correct answer.
    4. A weekly capacity review meeting

    Explanation

    An alarm evaluates a metric against a threshold over a configured period and triggers notification or automated action when the condition holds. The evaluation period is what makes the five-minute duration meaningful.

  98. Question 98Deployment, Provisioning, and Operations

    What is the most reliable way to keep a large fleet of instances patched without manual effort?

    1. Patch each host by hand during business hours
    2. Disable automatic updates and wait for reports of exploits
    3. Bake updated images and roll them out through the deployment pipeline Correct answer.
    4. Grant every engineer administrative access to patch ad hoc

    Explanation

    Rebuilding images with the patched software and rolling them out through the pipeline makes patching automated, testable, and uniform across the fleet. It also preserves the ability to roll back to the previous image.

  99. Question 99Deployment, Provisioning, and Operations

    Which practice makes a deployment reproducible from a known good build?

    1. Rebuilding the artifact from the latest branch at deploy time
    2. Copying files onto the server over an interactive session
    3. Allowing each environment to compile its own binaries
    4. Promoting the same versioned artifact through each environment Correct answer.

    Explanation

    Building once and promoting that identical artifact through environments guarantees that what was tested is what runs in production. Recompiling per environment can introduce differences that were never validated.

  100. Question 100Deployment, Provisioning, and Operations

    An auto-scaling group repeatedly launches and terminates instances as load fluctuates. Which setting should be tuned first?

    1. The instance type family
    2. The load balancing algorithm
    3. The scale-in cooldown period Correct answer.
    4. The number of availability zones

    Explanation

    Flapping usually means the group scales in before newly added capacity has had time to change the metric, so lengthening the cooldown stabilizes the decision loop. Cooldown settings govern how long the group waits before reacting again.

  101. Question 101Deployment, Provisioning, and Operations

    Why does a declarative provisioning tool report no changes when the template is applied twice with no edits?

    1. It records the run in a log and suppresses further applies
    2. It creates duplicate resources but hides them from the output
    3. It converges the live state to the declared state, which already matches Correct answer.
    4. It caches the first run and skips all subsequent executions

    Explanation

    Declarative tools compare the declared configuration with the observed live state on every run and only act on differences. Once the two match, the plan is empty and no changes are made.

  102. Question 102Deployment, Provisioning, and Operations

    A workload keeps a scaled-down copy of the full stack running in the recovery region and scales it up during a disaster. Which disaster recovery strategy is this?

    1. Warm standby Correct answer.
    2. Pilot light
    3. Backup and restore
    4. Multi-site active-active

    Explanation

    Warm standby maintains a functional but reduced-capacity copy of the entire stack in the recovery site, which is scaled up when failover occurs. Pilot light differs by keeping only a minimal core such as replicated data and provisioning the rest at failover time.

  103. Question 103Deployment, Provisioning, and Operations

    A container orchestrator is asked to run three replicas, but one node fails. What does the controller do?

    1. It waits for an operator to file a replacement request
    2. It reduces the desired count to match the running pods
    3. It reschedules the missing replica onto a healthy node Correct answer.
    4. It promotes the remaining replicas to a larger size

    Explanation

    The controller runs a reconciliation loop that continuously compares observed state with the declared replica count and schedules replacements where capacity exists. This self-healing behavior restores the declared state without human intervention.

  104. Question 104Deployment, Provisioning, and Operations

    Before switching production traffic to a new environment by changing a DNS record, what should be done to reduce the cutover window?

    1. Lower the record's time to live in advance of the change Correct answer.
    2. Disable health checks on both environments
    3. Point the record at the old environment's load balancer
    4. Increase the record's time to live well in advance

    Explanation

    Resolvers cache DNS answers for the duration of the TTL, so lowering it ahead of time means stale answers expire quickly and clients pick up the new target. Raising it has the opposite effect and prolongs the transition.

  105. Question 105Deployment, Provisioning, and Operations

    A CI/CD pipeline needs to create and modify cloud resources. Which access model should it use?

    1. A shared administrator account used by all pipelines
    2. Long-lived access keys stored in the pipeline's repository
    3. A dedicated role scoped to the resources the pipeline manages Correct answer.
    4. The personal credentials of the engineer who wrote the pipeline

    Explanation

    Automation should assume a dedicated identity granted only the permissions needed for the resources it manages, following least privilege. This limits blast radius and produces an auditable record of automated changes.

  106. Question 106Deployment, Provisioning, and Operations

    In a blue/green deployment, production traffic is switched from the blue environment to the green environment. What makes rollback fast if the new release proves faulty?

    1. The green environment is rebuilt from the last available backup.
    2. Traffic is switched back to the still-running blue environment. Correct answer.
    3. The pipeline re-runs the full automated test suite against green.
    4. Each instance in green is patched in place back to the previous version.

    Explanation

    Blue/green keeps the previous environment intact and ready, so reverting is a traffic switch rather than a rebuild or redeploy and completes in seconds. The old environment is only torn down after the new one has proven stable.

  107. Question 107Deployment, Provisioning, and Operations

    A pipeline deploys a new build and then runs a small set of critical-path checks against the running service before marking the release successful. What is the purpose of these checks?

    1. To verify that the source code compiles without warnings.
    2. To measure how much of the source code the unit tests exercise.
    3. To confirm the deployed service works in its real environment. Correct answer.
    4. To benchmark the service against comparable products.

    Explanation

    Smoke tests exercise the deployed artifact in the target environment, catching problems with configuration, dependencies, and connectivity that unit tests cannot see. They act as a gate before the release is declared good.

  108. Question 108Deployment, Provisioning, and Operations

    A workload's response time degrades as request volume grows. The team responds by adding more instances of the same size behind the load balancer. Which scaling approach is this?

    1. Horizontal scaling Correct answer.
    2. Elastic caching
    3. Vertical scaling
    4. Data sharding

    Explanation

    Horizontal scaling adds more nodes of the same size and spreads load across them. Vertical scaling instead means moving to a larger, more powerful node.

  109. Question 109Deployment, Provisioning, and Operations

    Two engineers run an infrastructure-as-code apply against the same stack at the same time. Which control prevents them from corrupting the shared state?

    1. Keeping a nightly backup copy of the state file.
    2. Issuing both engineers a read-only API token.
    3. Requiring each engineer to work on a separate branch.
    4. Holding a state lock for the duration of each apply. Correct answer.

    Explanation

    A state lock allows only one operation to modify the state at a time, serializing concurrent applies so they cannot overwrite each other. Backups and branches do not stop simultaneous writes.

  110. Question 110Deployment, Provisioning, and Operations

    An instance behind a load balancer passes its operating system status check, but its application process has crashed. Which configuration causes the load balancer to stop sending traffic to that instance?

    1. A larger instance type in the target group.
    2. A longer connection idle timeout on the load balancer.
    3. Application-level health checks polled by the load balancer. Correct answer.
    4. Sticky sessions enabled on the target group.

    Explanation

    An application-level health check probes the service itself, so a crashed process fails the check and the target is removed from rotation even though the host is healthy. Host-level status checks only report on the machine.

  111. Question 111Deployment, Provisioning, and Operations

    A team deploys containers using the tag latest. What is the main risk of this practice?

    1. It prevents the registry from storing more than one version.
    2. It disables the health check defined in the image.
    3. It compresses the image and slows container startup.
    4. The same tag can point to different image content over time. Correct answer.

    Explanation

    Tags are mutable references, so latest can resolve to a different image on each pull, making the running version ambiguous and rollbacks unreliable. Pinning to a unique tag or digest keeps deployments reproducible.

  112. Question 112Deployment, Provisioning, and Operations

    Which practice best reduces the chance that a release works in staging but fails in production?

    1. Freezing the staging environment permanently after one successful deployment so it never changes again.
    2. Giving production much larger capacity than staging.
    3. Building both environments from the same templates and baselines. Correct answer.
    4. Allowing engineers to hotfix production by hand when needed.

    Explanation

    Environment parity means staging and production share the same provisioning templates, runtime versions, and configuration baselines, so differences that cause release failures are exposed before production.

  113. Question 113Deployment, Provisioning, and Operations

    Instances are spread across three regions and engineers must search logs from all of them in one place. What should be implemented?

    1. A log viewer installed on each engineer's workstation.
    2. Forwarding logs to a central log aggregation service. Correct answer.
    3. Per-region logins that engineers check one region at a time.
    4. Hourly log rotation on each instance with local retention.

    Explanation

    Shipping logs to a central aggregation service gives one searchable index across regions and survives the loss of any single instance. Local files and per-region access cannot provide a single query point.

  114. Question 114Deployment, Provisioning, and Operations

    A database credential used by a running service must be rotated with no downtime. Which sequence is correct?

    1. Create the new credential, deploy it to the service, verify it works, then revoke the old one. Correct answer.
    2. Revoke the old credential first, then create and deploy the new one immediately.
    3. Create the new credential and commit it to the application repository for the next build.
    4. Restart all service instances first, then change the credential in the database.

    Explanation

    Rotation without downtime relies on an overlap window in which both credentials are valid, so instances can pick up the new secret while the old one still authenticates. The old credential is revoked only after verification.

  115. Question 115Deployment, Provisioning, and Operations

    A team must rename a column in a live database while old and new application code run at the same time during a rolling deployment. Which approach avoids errors?

    1. Drop the old column first so the schema is clean before the code change ships.
    2. Add the new column, write to both, backfill, switch reads, then drop the old column. Correct answer.
    3. Rename the column in a single migration and deploy the code in the same step.
    4. Lock the table exclusively for the entire deployment so that no writes occur during the migration.

    Explanation

    Expanding first and contracting later keeps the schema compatible with both code versions throughout the rollout. Each step is independently safe, and the old column is removed only once no running code references it.

  116. Question 116Deployment, Provisioning, and Operations

    A team wants to deploy code to production while keeping a new feature hidden from users until it is enabled later. Which mechanism fits?

    1. A long-lived release branch merged only at launch time.
    2. A feature flag evaluated at runtime. Correct answer.
    3. A separate production environment for each feature.
    4. A database read replica dedicated to the feature.

    Explanation

    A feature flag separates deployment from release: code ships to production but the feature stays off until the flag is turned on. This also allows quick disabling without a redeploy.

  117. Question 117Networking and Connectivity

    Which OSI model layer is responsible for logical addressing and for choosing a path between networks?

    1. Transport layer
    2. Application layer
    3. Data link layer
    4. Network layer Correct answer.

    Explanation

    The network layer (layer 3) handles logical addressing and path selection between networks, which is where IP operates. The data link layer deals with physical addressing on a single link, while the transport layer handles end-to-end delivery between processes.

  118. Question 118Networking and Connectivity

    Which transport protocol establishes a connection and guarantees ordered, reliable delivery of data?

    1. TCP Correct answer.
    2. SNMP
    3. UDP
    4. ICMP

    Explanation

    TCP performs a handshake before data transfer, acknowledges received segments, and retransmits lost data, so delivery is reliable and ordered. UDP sends datagrams without a connection or delivery guarantee.

  119. Question 119Networking and Connectivity

    Which IPv4 address range is reserved for private use inside an organization's network?

    1. 169.254.0.0/16
    2. 224.0.0.0/4
    3. 172.16.0.0/12 Correct answer.
    4. 192.0.2.0/24

    Explanation

    The private ranges defined for internal use are 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16; these are not routed on the public internet. The other ranges serve documentation, link-local, and multicast purposes.

  120. Question 120Networking and Connectivity

    Which TCP port is used by HTTPS by default?

    1. 21
    2. 22
    3. 443 Correct answer.
    4. 80

    Explanation

    HTTPS uses TCP port 443, which carries HTTP over a TLS session. Port 80 is plain HTTP, port 21 is FTP control, and port 22 is SSH.

  121. Question 121Networking and Connectivity

    What is the primary function of DNS in an IP network?

    1. Route packets between autonomous systems
    2. Encrypt traffic between clients and servers
    3. Map hostnames to IP addresses Correct answer.
    4. Automatically assign IP addresses to hosts

    Explanation

    DNS is a distributed naming system that resolves human-readable names into the IP addresses that hosts actually use. Address assignment is handled by DHCP, encryption by TLS, and inter-domain routing by BGP.

  122. Question 122Networking and Connectivity

    What does a DHCP server provide to a host that joins a network?

    1. Encrypted tunnels between two sites
    2. Name resolution for external domains
    3. Automatic assignment of IP configuration Correct answer.
    4. Load distribution across web servers

    Explanation

    DHCP leases an IP address plus supporting parameters such as the subnet mask, default gateway, and DNS servers, removing the need for manual configuration. Name resolution, tunneling, and load distribution are separate services.

  123. Question 123Networking and Connectivity

    How many bits long is an IPv6 address?

    1. 32
    2. 128 Correct answer.
    3. 64
    4. 256

    Explanation

    IPv6 addresses are 128 bits, usually written as eight groups of four hexadecimal digits. IPv4 addresses are 32 bits, and the 64-bit figure refers to the typical interface identifier portion.

  124. Question 124Networking and Connectivity

    How many usable host addresses are available in a /28 IPv4 subnet?

    1. 30
    2. 14 Correct answer.
    3. 16
    4. 32

    Explanation

    A /28 leaves 4 host bits, giving 2^4 = 16 total addresses. Subtracting the network and broadcast addresses leaves 14 usable host addresses.

  125. Question 125Networking and Connectivity

    Which statement describes how a cloud security group handles traffic for an attached instance?

    1. It supports explicit deny rules that override allow rules
    2. It is stateless, so each direction needs its own rule
    3. It evaluates numbered rules in order until a match is found
    4. It is stateful, so reply traffic is permitted automatically Correct answer.

    Explanation

    Security groups are stateful: when an inbound connection is allowed, the response traffic is tracked and permitted without a matching outbound rule. Numbered, ordered evaluation and explicit deny rules are characteristics of network ACLs.

  126. Question 126Networking and Connectivity

    At which point in a virtual private cloud is a network ACL enforced?

    1. On the load balancer's listener
    2. At the subnet boundary Correct answer.
    3. On the instance's virtual network interface
    4. Inside the guest operating system

    Explanation

    A network ACL is a stateless filter attached to a subnet that evaluates traffic entering and leaving that subnet. Instance-level filtering is done by security groups, and guest-level filtering is done by the host's own firewall.

  127. Question 127Networking and Connectivity

    Instances in a private subnet must download updates from the internet but must not be reachable from it. What should be configured?

    1. An internet gateway attached directly to the private subnet
    2. A public load balancer in front of the instances
    3. A NAT gateway in a public subnet with a route Correct answer.
    4. A VPN gateway with a static public IP address

    Explanation

    A NAT gateway performs source address translation so private instances can initiate outbound connections while remaining unreachable from the internet. The private subnet's route table sends internet-bound traffic to the NAT gateway, which resides in a public subnet.

  128. Question 128Networking and Connectivity

    Requests for /images must reach an image fleet and requests for /api must reach an API fleet. Which capability is required?

    1. A layer 4 network load balancer with TCP listeners
    2. DNS round-robin across both fleets
    3. A layer 7 load balancer with path-based rules Correct answer.
    4. A layer 3 router with static routes

    Explanation

    Routing on the request URL path requires inspecting the HTTP request, which is application-layer (layer 7) behavior. Layer 4 balancers only see IP addresses and ports, and static routes operate below the HTTP level entirely.

  129. Question 129Networking and Connectivity

    A website with a single origin server serves images slowly to users on other continents. Which change reduces latency the most?

    1. Cache content at edge locations close to users Correct answer.
    2. Resize the origin server to a larger instance
    3. Open a site-to-site VPN to each user region
    4. Add another availability zone in the origin region

    Explanation

    Content delivery networks keep copies of static assets at points of presence near users, shortening the network distance each request travels. Adding capacity or zones at the origin does not reduce the round-trip distance to distant clients.

  130. Question 130Networking and Connectivity

    VPC A is peered with VPC B, and VPC B is peered with VPC C. What happens to traffic sent from A to C by default?

    1. It succeeds if B disables route propagation
    2. It is forwarded automatically by B's routing tables
    3. It is dropped, because peering is not transitive Correct answer.
    4. It succeeds only through a NAT instance in B

    Explanation

    VPC peering creates a one-to-one relationship between two VPCs and is not transitive, so A cannot reach C through B. Reaching C requires a direct peering, a transit gateway, or another routing hub.

  131. Question 131Networking and Connectivity

    A route table contains 10.0.0.0/16 pointing to local and 10.0.1.0/24 pointing to a firewall appliance. Where is traffic destined for 10.0.1.5 sent?

    1. To the appliance, because /24 is the longest match Correct answer.
    2. To local, because /16 covers more addresses
    3. To neither, because overlapping routes are rejected
    4. To both targets, using equal-cost multipath

    Explanation

    Routers select the most specific matching route, meaning the longest prefix that contains the destination address. Here 10.0.1.0/24 is more specific than 10.0.0.0/16, so the appliance receives the traffic.

  132. Question 132Networking and Connectivity

    An application transfers little data but makes many small sequential requests, and the average round-trip time to the server is 250 ms. Which factor dominates the slow response?

    1. Latency Correct answer.
    2. Jitter
    3. Packet loss
    4. Bandwidth

    Explanation

    Each sequential request must wait for a full round trip, so the delay compounds with propagation time rather than with transfer size. High bandwidth does not help when the workload is dominated by round-trip delays.

  133. Question 133Networking and Connectivity

    What is the effect of lowering the time-to-live value on a DNS record?

    1. Resolvers keep cached answers for a longer period
    2. Changes propagate faster, at the cost of more DNS queries Correct answer.
    3. The record is served only to authoritative queries
    4. Changes propagate slower, but queries become cheaper

    Explanation

    The TTL is the maximum time a resolver may cache the record, so a shorter TTL means stale data clears sooner after a change. The trade-off is more frequent lookups reaching authoritative servers.

  134. Question 134Networking and Connectivity

    An engineer must divide 192.168.1.0/24 into four equal-sized subnets. Which prefix length results?

    1. /26 Correct answer.
    2. /28
    3. /30
    4. /23

    Explanation

    Splitting a /24 into four subnets requires two additional bits, moving the prefix from 24 to 26. Each resulting /26 contains 64 addresses.

  135. Question 135Networking and Connectivity

    A company wants to terminate TLS in one place and forward requests to backend servers without exposing them. Which component performs this?

    1. A layer 4 network load balancer
    2. A reverse proxy in front of the servers Correct answer.
    3. A NAT gateway in the public subnet
    4. A forward proxy installed on each client

    Explanation

    A reverse proxy accepts client connections on behalf of backend servers, so it can terminate TLS and forward requests without revealing the servers' addresses. Forward proxies act on behalf of clients, while layer 4 balancers do not have to terminate TLS.

  136. Question 136Networking and Connectivity

    An enterprise has a primary dedicated private link and a backup IPsec VPN to the same cloud region. Which mechanism enables automatic failover between them?

    1. A virtual router redundancy protocol on the gateway
    2. BGP with route preference on the advertised prefixes Correct answer.
    3. Static routes with equal administrative distance
    4. DNS round-robin across both paths

    Explanation

    BGP dynamically exchanges reachability information, so a failed path is withdrawn and traffic moves to the surviving path without manual intervention. Attributes on the advertised prefixes let the network prefer the dedicated link while it is healthy.

  137. Question 137Networking and Connectivity

    An IPsec tunnel limits the usable MTU to 1400 bytes, and applications sending 1500-byte packets with the DF bit set hang. What is the appropriate remedy?

    1. Move the application from TCP to UDP
    2. Reduce the TCP MSS or enable path MTU discovery Correct answer.
    3. Clear the DF bit on all core routers
    4. Raise the tunnel MTU to 9000 bytes

    Explanation

    Encapsulation consumes header space, so payloads must fit within the smaller tunnel MTU. Clamping TCP MSS or relying on path MTU discovery lets endpoints size segments correctly and avoid black-holed packets.

  138. Question 138Networking and Connectivity

    How does an overlay network deliver traffic across a routed underlay?

    1. It assigns public IP addresses to every virtual machine
    2. It replaces IP routing with static MAC address forwarding
    3. It requires all endpoints to share a single broadcast domain
    4. It encapsulates frames inside IP packets sent over the underlay Correct answer.

    Explanation

    Overlay technologies such as VXLAN take a layer 2 frame and wrap it in an IP packet, so it can cross a routed layer 3 fabric. This decouples the virtual topology from the physical topology and avoids stretching broadcast domains across sites.

  139. Question 139Networking and Connectivity

    How does anycast reduce latency for a globally distributed service?

    1. A single address is announced from many locations Correct answer.
    2. It resolves the hostname to a different address per query
    3. It compresses payloads before they leave the origin server
    4. It caches responses in each user's local browser

    Explanation

    With anycast, multiple sites advertise the same address into the routing system, so each user's traffic is delivered to a topologically near instance. The client configuration does not change, and routing adapts if a site becomes unreachable.

  140. Question 140Networking and Connectivity

    An organization has 40 VPCs that all need to communicate, and full-mesh peering has become unmanageable. Which design scales best?

    1. A single VPC that hosts every workload
    2. One NAT gateway shared by all VPCs
    3. Static routes exchanged manually between each pair
    4. A hub-and-spoke transit gateway with VPC attachments Correct answer.

    Explanation

    A transit gateway acts as a regional hub, so each VPC attaches once instead of peering with every other VPC, reducing the connection count from a quadratic mesh to a linear set of attachments. Route tables on the hub then control which VPCs can reach each other.

  141. Question 141Data Management and Storage

    Which storage type presents raw logical volumes that an operating system partitions and formats with its own file system?

    1. Block storage Correct answer.
    2. File storage
    3. Archive storage
    4. Object storage

    Explanation

    Block storage exposes fixed-size blocks addressed by number over protocols such as iSCSI or Fibre Channel, and the attached host owns the file system. Object and file storage instead manage the namespace themselves, so the host does not format them.

  142. Question 142Data Management and Storage

    In object storage, what uniquely identifies an object inside a bucket?

    1. A LUN identifier
    2. An inode number
    3. A mount point path
    4. An object key Correct answer.

    Explanation

    Each object is addressed by a key that is unique within its bucket, and clients retrieve it with bucket-plus-key over an HTTP API. The flat namespace means there are no directory inodes or block device identifiers involved.

  143. Question 143Data Management and Storage

    What does the 3-2-1 backup rule recommend?

    1. One copy encrypted and replicated twice inside the same array
    2. Three copies of data on three separate cloud regions with no offline copy
    3. Three copies across two media types with one copy offsite Correct answer.
    4. Two copies of data on one media type kept in the same rack

    Explanation

    The rule keeps three total copies of data on two different media types, with at least one copy stored offsite or offline. This diversification protects against a single media failure, site outage, or destructive event affecting all copies.

  144. Question 144Data Management and Storage

    Which metric defines the maximum acceptable amount of data loss, measured in time, after a disruption?

    1. MTTR
    2. RPO Correct answer.
    3. SLA
    4. RTO

    Explanation

    The recovery point objective states how far back in time recovery may go, which drives backup and replication frequency. The recovery time objective, by contrast, bounds how long restoration may take.

  145. Question 145Data Management and Storage

    Encrypting a database volume so that data cannot be read from stolen physical disks addresses which protection goal?

    1. Data at rest Correct answer.
    2. Data in use
    3. Data in motion
    4. Data in transit

    Explanation

    Data at rest is data stored on media, and volume-level encryption protects it when the physical device leaves the owner's control. Transit and motion encryption protect data on the wire, while in-use protection covers data in memory or during processing.

  146. Question 146Data Management and Storage

    Which practice moves infrequently accessed data to lower-cost media while keeping active data on faster storage?

    1. Sharding
    2. Lifecycle tiering Correct answer.
    3. Replication
    4. Erasure striping

    Explanation

    Tiering places data on media matched to its access frequency, so hot data stays on fast storage and cold data migrates to cheaper tiers. Sharding and replication change distribution or copy count but not cost-by-access-pattern.

  147. Question 147Data Management and Storage

    What does a storage snapshot capture?

    1. An offline tape copy shipped to a remote vault
    2. A live replica synchronized to a second region
    3. A point-in-time copy of a volume Correct answer.
    4. A checksum used to verify block integrity

    Explanation

    A snapshot records the state of a volume at a moment in time so it can later be restored or cloned, and it is often implemented with copy-on-write or redirect-on-write. It is not automatically a separate-site copy or an integrity check.

  148. Question 148Data Management and Storage

    A four-disk RAID 5 array loses one member disk. What is the resulting state?

    1. It continues in degraded mode until the disk is rebuilt Correct answer.
    2. It goes offline and all array data is lost
    3. It keeps full redundancy because two parity copies exist
    4. It becomes read-only until an administrator rebuilds it

    Explanation

    RAID 5 distributes a single parity block per stripe, so one missing member can be reconstructed on the fly and the array keeps serving I/O with reduced performance and no redundancy. Replacing the disk triggers a rebuild that restores protection.

  149. Question 149Data Management and Storage

    Which RAID level stripes data across mirrored pairs and can tolerate one disk failure in each pair?

    1. RAID 10 Correct answer.
    2. RAID 0
    3. RAID 6
    4. RAID 5

    Explanation

    RAID 10 combines mirroring with striping, so each mirror pair survives the loss of one member while the stripe set provides performance. RAID 5 and 6 use distributed parity rather than mirroring, and RAID 0 has no redundancy at all.

  150. Question 150Data Management and Storage

    What characterizes asynchronous replication between a primary and a secondary site?

    1. Every write must be committed at both sites before it is acknowledged, so no data is lost
    2. Writes are acknowledged locally and shipped later, so recent writes can be lost Correct answer.
    3. Replication pauses automatically whenever inter-site latency exceeds a threshold
    4. The secondary site is refreshed only during scheduled nightly backup windows

    Explanation

    With asynchronous replication the primary acknowledges writes without waiting for the remote site, giving lower write latency but a non-zero recovery point objective. Only synchronous replication guarantees zero loss of acknowledged writes at the cost of latency.

  151. Question 151Data Management and Storage

    A client writes a value and then immediately reads the same key in the same session. A store that always returns the newly written value is providing which guarantee?

    1. Read-your-writes consistency Correct answer.
    2. Eventual consistency with a bounded staleness window
    3. Causal consistency across unrelated keys
    4. Strong global serializability across all regions

    Explanation

    Read-your-writes consistency guarantees that a session observes its own prior writes, which is a per-session guarantee. It is weaker than global serializability and stronger than eventual consistency.

  152. Question 152Data Management and Storage

    What does data deduplication accomplish in a backup repository?

    1. Each block is compressed with a lossless algorithm before it is written
    2. Blocks are encrypted so identical plaintext yields different ciphertext
    3. Duplicate blocks are stored once and referenced thereafter Correct answer.
    4. Blocks are spread across disks to increase read throughput

    Explanation

    Deduplication identifies repeated chunks and stores a single instance plus references, sharply reducing consumed capacity across backup generations. It is distinct from compression, encryption, and striping, which solve different problems.

  153. Question 153Data Management and Storage

    A workload issues a high volume of small random reads. Which storage characteristic most directly constrains its performance?

    1. Capacity in gigabytes
    2. Number of namespace buckets
    3. Sequential bandwidth in MB/s
    4. IOPS Correct answer.

    Explanation

    Small random I/O is dominated by the number of operations completed per second, which is measured in IOPS. Sequential bandwidth matters for large streaming transfers rather than scattered small reads.

  154. Question 154Data Management and Storage

    What does thin provisioning mean in an enterprise storage array?

    1. Each volume is pinned to a dedicated physical disk for isolation
    2. Volumes are automatically compressed to half of their nominal size
    3. Physical capacity is fully pre-allocated up front so that writes can never fail
    4. Logical capacity is presented beyond physical allocation and consumed as data is written Correct answer.

    Explanation

    Thin provisioning presents a larger logical size than the capacity currently allocated, drawing from a shared pool as blocks are actually written. This raises utilization but requires monitoring so the pool does not run out of space.

  155. Question 155Data Management and Storage

    A bucket lifecycle policy transitions objects to cold storage after 90 days and deletes them after 365 days. What is the primary business purpose?

    1. Increase durability by replicating objects to more regions
    2. Lower cost while still meeting retention requirements Correct answer.
    3. Improve read latency for frequently accessed objects
    4. Enforce encryption of objects at rest with customer keys

    Explanation

    Lifecycle rules align storage class and deletion timing with how long data must be retained, which minimizes spend on inactive data. Durability, latency, and encryption are governed by other configuration settings.

  156. Question 156Data Management and Storage

    How does envelope encryption protect objects in cloud storage?

    1. A data key encrypts data; a KMS master key wraps the data key Correct answer.
    2. A per-object key is stored in plaintext alongside the object
    3. TLS session keys are reused as the storage encryption keys
    4. One long-lived master key encrypts both data and transport traffic

    Explanation

    Envelope encryption uses a unique data key per object and protects that key with a master key held in a key management service, so the master key never leaves the service. Rotating or revoking the master key limits exposure without re-encrypting every object directly.

  157. Question 157Data Management and Storage

    Regulations require that customer records never leave the country of origin. Which control most directly satisfies this?

    1. Encrypting every record with keys managed by the cloud provider
    2. Enabling versioning and object lock on the bucket
    3. Keeping storage and replication inside that country Correct answer.
    4. Replicating records to three continents for durability

    Explanation

    Data residency is a placement requirement, so the effective control is restricting the regions used for primary storage, backups, and replicas. Encryption, versioning, and multi-continent replication do not change where the data physically resides.

  158. Question 158Data Management and Storage

    Which statement correctly distinguishes NAS from SAN?

    1. NAS serves blocks over Ethernet; SAN serves files over Fibre Channel
    2. NAS requires Fibre Channel; SAN runs only over standard Ethernet
    3. NAS serves files over NFS or SMB; SAN serves blocks over a dedicated network Correct answer.
    4. NAS is used only for backup targets; SAN is used only for databases

    Explanation

    NAS presents a shared file system over protocols such as NFS or SMB, while SAN delivers block devices to hosts, traditionally over Fibre Channel or iSCSI. The access model, not the workload type, is the defining difference.

  159. Question 159Data Management and Storage

    Monitoring shows a 100 TB storage pool 60% full with usage growing about 8% of the pool per month. Which action is most appropriate?

    1. Cut retention to zero days so that growth stops immediately
    2. Wait until utilization reaches 95% before investigating, since headroom exists
    3. Forecast exhaustion in about six to eight months and plan now Correct answer.
    4. Buy another 100 TB immediately without further analysis

    Explanation

    Trend-based projection converts a growth rate into a runway estimate, here roughly 40 TB of free space against about 8 TB of monthly growth. Planning expansion, tiering, or archiving before thresholds are breached avoids an outage.

  160. Question 160Data Management and Storage

    A storage system applies a 10-of-14 erasure coding scheme to each object. What overhead and fault tolerance does this provide?

    1. 40% overhead with tolerance for four simultaneous failures Correct answer.
    2. 100% overhead with tolerance for up to seven failures
    3. Zero overhead with tolerance for one failure because blocks are mirrored
    4. 14% overhead with tolerance for up to ten failures

    Explanation

    Four parity fragments are added to ten data fragments, so usable payload is 10/14 of stored bytes, which is roughly 40 percent overhead. Any four fragments may be lost and the original object can still be reconstructed.

  161. Question 161Data Management and Storage

    A globally distributed key-value store accepts writes in any region and replicates them asynchronously. Two clients in different regions update the same key at nearly the same time. What must the application be prepared to handle?

    1. Blocking of the second write until the first has fully replicated
    2. Automatic rollback of the later write with an error returned to the client
    3. Strict serial ordering of both writes as if they hit a single node
    4. Conflicting versions the application must resolve, e.g. last-write-wins Correct answer.

    Explanation

    Multi-region asynchronous writes with eventual consistency can produce divergent versions of the same key, so the system needs a conflict-resolution rule such as last-write-wins, version vectors, or application merge logic. The trade-off for this availability is the loss of a single global serial order.

  162. Question 162Data Management and Storage

    Ransomware encrypted production data and also reached the mounted backups. Which control best ensures future recoverability?

    1. Backups encrypted with a key stored on the production file server
    2. A backup catalog on a domain-joined server using the same admin account
    3. Daily snapshots kept on the same SAN as the production volumes
    4. Immutable off-network copies with object lock and separate credentials Correct answer.

    Explanation

    Recoverability after ransomware depends on copies that the attacker cannot delete or encrypt, which requires immutability, network isolation, and credentials outside the compromised domain. Snapshots and catalogs sharing the production trust boundary can be destroyed along with the data.

  163. Question 163Data Management and Storage

    One tenant's workload saturates a shared storage array and degrades latency for all other tenants. Which mechanism most directly addresses this?

    1. Moving the catalog database to the same LUN
    2. Per-tenant QoS limits on IOPS and bandwidth Correct answer.
    3. Enabling deduplication on all volumes
    4. Adding raw capacity to the array

    Explanation

    Quality-of-service limits cap how many operations and how much bandwidth a tenant can consume, preventing one workload from monopolizing shared resources. Capacity and data-reduction features do not constrain contention for I/O service.

  164. Question 164Data Management and Storage

    A 200 TB dataset must reach the cloud within two weeks, but the available link provides about 1 Gbps. Which approach is most appropriate?

    1. Split the data into 200 parallel streams to exceed line rate
    2. Ship encrypted disks or use an offline transfer appliance Correct answer.
    3. Throttle the transfer to off-peak hours to raise throughput
    4. Compress the data and retry the online transfer, since the link is sufficient

    Explanation

    At roughly 1 Gbps the theoretical ceiling is about 10 TB per day, and practical throughput with protocol overhead is considerably lower, so 200 TB will not fit the two-week window. Physical shipment or a dedicated transfer appliance bypasses the bandwidth constraint, after which only incremental changes are transferred online.

  165. Question 165Automation, DevOps, and Infrastructure as Code

    A team runs its provisioning pipeline twice against unchanged source code, and the resulting infrastructure is identical both times, with no duplicate resources created. Which property of infrastructure as code does this demonstrate?

    1. Observability
    2. Idempotence Correct answer.
    3. Immutability
    4. Elasticity

    Explanation

    Idempotence means applying the same configuration repeatedly converges on the same end state rather than creating additional or duplicated resources. It is a core reason IaC tools are safe to re-run on every pipeline execution. Immutability, elasticity and observability describe other properties of infrastructure and delivery systems.

  166. Question 166Automation, DevOps, and Infrastructure as Code

    Which practice is the clearest example of continuous delivery rather than continuous deployment?

    1. Every merge to the main branch produces a tested, releasable artifact that goes live after a manual approval. Correct answer.
    2. Developers merge their work to the main branch at least once per day.
    3. Every merge to the main branch is pushed to production automatically with no human action.
    4. All infrastructure is rebuilt from code on every release.

    Explanation

    Continuous delivery keeps every change in a deployable state but leaves the decision to release to a human approval step. Continuous deployment removes that gate and releases automatically. Frequent merging and rebuilding infrastructure describe other practices.

  167. Question 167Automation, DevOps, and Infrastructure as Code

    What is the primary purpose of running terraform plan before terraform apply?

    1. It downloads and installs the providers required by the configuration.
    2. It writes the current real-world infrastructure into the state file.
    3. It checks configuration syntax and nothing else.
    4. It previews the changes that apply would make to the managed infrastructure. Correct answer.

    Explanation

    A plan compares configuration, state and (by default) real resources to show the create, update and destroy actions that apply would perform. Reviewing it is the main safety gate in a Terraform pipeline. Provider installation is handled by terraform init, and syntax checking is handled by terraform validate.

  168. Question 168Automation, DevOps, and Infrastructure as Code

    In Terraform, which component translates resource definitions into API calls for a specific platform such as a cloud provider?

    1. A backend
    2. A module
    3. A provider Correct answer.
    4. A workspace

    Explanation

    Providers are plugins that implement the API calls for a given platform and map Terraform resources onto that platform's objects. Modules package configuration, workspaces separate state instances, and backends store and lock state.

  169. Question 169Automation, DevOps, and Infrastructure as Code

    What is the main benefit of wrapping a new feature in a feature flag?

    1. It removes the need for automated tests.
    2. It guarantees the feature has no performance impact.
    3. It eliminates the need for a rollback plan.
    4. Code deployment is decoupled from feature release. Correct answer.

    Explanation

    A feature flag lets code ship to production while the behaviour stays off until it is deliberately enabled for chosen users. This separates deployment from release and allows rapid disabling without redeploying. It does not replace testing, rollback planning or performance validation.

  170. Question 170Automation, DevOps, and Infrastructure as Code

    A team describes the desired end state of its network in code and lets the tool determine which API calls are needed to reach it. Which style of automation is this?

    1. Procedural orchestration
    2. Imperative scripting
    3. Declarative configuration Correct answer.
    4. Manual provisioning

    Explanation

    Declarative configuration states what the result should be and lets the tool compute the steps, which is how Terraform-style tools operate. Imperative approaches spell out each step in order. The other options either describe step-by-step execution or no automation at all.

  171. Question 171Automation, DevOps, and Infrastructure as Code

    Why must a Terraform state file be available when applying a configuration that already manages live resources?

    1. It caches the output of the last plan for auditing purposes.
    2. It holds the version history of the configuration files.
    3. It maps configuration resources to the real objects they created. Correct answer.
    4. It stores the provider credentials used for authentication.

    Explanation

    State records the binding between each resource address in configuration and the real object's identifier, so the tool can update or delete the correct objects instead of creating duplicates. Credentials are supplied through providers and environment configuration, and version history lives in source control.

  172. Question 172Automation, DevOps, and Infrastructure as Code

    An engineer edits a security group rule in the cloud console. On the next pipeline run, Terraform proposes to revert that change. What does this illustrate?

    1. A module dependency cycle
    2. A state locking failure
    3. Provider version pinning taking effect
    4. Configuration drift between code and real infrastructure Correct answer.

    Explanation

    Drift occurs when real infrastructure diverges from what the configuration and state describe. Planning refreshes the real objects and then proposes actions that bring them back to the declared state. Locking, version pinning and dependency cycles produce different symptoms.

  173. Question 173Automation, DevOps, and Infrastructure as Code

    Which characteristic distinguishes a blue/green deployment?

    1. Users are split by geography and compared over several weeks.
    2. A small percentage of users receive the new version first.
    3. The new version is deployed only after the old one is removed.
    4. All traffic is switched to a verified identical environment. Correct answer.

    Explanation

    Blue/green keeps two parallel environments and switches traffic over once the new one passes verification, enabling fast rollback by switching back. Gradual exposure to a subset of users is canary, and removing the old version before deploying the new one is a downtime-prone cutover.

  174. Question 174Automation, DevOps, and Infrastructure as Code

    Which statement best describes a canary release strategy?

    1. Sending a small slice of traffic to the new version while monitoring error rates. Correct answer.
    2. Rebuilding every server from an image on each deployment.
    3. Running two complete environments and switching DNS records between them.
    4. Routing all user traffic to the new version immediately and rolling back if errors appear.

    Explanation

    A canary directs a limited share of traffic to the new version so that health signals can be evaluated before widening exposure. The remaining options describe full cutover, blue/green, and immutable redeployment respectively.

  175. Question 175Automation, DevOps, and Infrastructure as Code

    A platform team wants the same VPC, subnet and firewall pattern reused across 30 services, with per-service values supplied by each caller. Which Terraform feature fits best?

    1. Data sources
    2. Provisioners
    3. Modules driven by input variables Correct answer.
    4. Workspaces

    Explanation

    Modules package reusable configuration, and input variables let each caller pass its own values into that shared pattern. Workspaces separate state instances, provisioners run scripts on resources, and data sources read existing information.

  176. Question 176Automation, DevOps, and Infrastructure as Code

    Infrastructure was created manually and must now be managed by Terraform without being recreated. Which approach applies?

    1. Write matching configuration and run terraform import. Correct answer.
    2. Run terraform destroy and then terraform apply.
    3. Add a lifecycle block with create_before_destroy.
    4. Delete the state file and re-run terraform apply.

    Explanation

    terraform import associates an existing real object with a resource address in state, after which the configuration must describe that object so plans are clean. Destroying and re-creating causes an outage, and an empty state would make Terraform try to create duplicates.

  177. Question 177Automation, DevOps, and Infrastructure as Code

    Which statement best distinguishes configuration management from provisioning?

    1. Provisioning installs and updates packages on servers; configuration management creates virtual networks.
    2. Configuration management always replaces servers instead of changing them.
    3. Both are used only to build container images.
    4. Provisioning creates infrastructure; configuration management installs and maintains software on it. Correct answer.

    Explanation

    Provisioning is about creating the infrastructure itself, such as networks, compute instances and storage, while configuration management keeps the software and settings on those machines in the desired state. They are complementary stages rather than the same activity.

  178. Question 178Automation, DevOps, and Infrastructure as Code

    How should a CI/CD pipeline obtain the cloud credentials it needs for a deployment?

    1. Commit an encrypted credential file to the shared repository for the pipeline.
    2. Store the keys in the container image as build arguments.
    3. Use short-lived credentials obtained at run time via OIDC federation. Correct answer.
    4. Embed long-lived access keys directly in the pipeline configuration file.

    Explanation

    Short-lived credentials issued at run time through identity federation or a secrets manager limit exposure if a pipeline is compromised. Long-lived keys in repositories, configuration or images are persistently exploitable and hard to rotate.

  179. Question 179Automation, DevOps, and Infrastructure as Code

    Deployment manifests that reference the :latest image tag create which operational risk?

    1. Autoscaling stops working for the workload.
    2. The registry rejects unsigned images.
    3. Images can no longer be pulled from the registry.
    4. Different hosts may run different versions under the same tag. Correct answer.

    Explanation

    A mutable tag such as latest can point to different image digests over time, so nodes that pull at different moments may run inconsistent code and rollbacks become ambiguous. Immutable, uniquely versioned tags keep deployments reproducible.

  180. Question 180Automation, DevOps, and Infrastructure as Code

    Which principle is central to GitOps?

    1. The running cluster is the source of truth and Git is updated from it.
    2. Git holds the desired state and an in-cluster agent reconciles toward it. Correct answer.
    3. Deployments are triggered only by messages in a chat channel.
    4. Engineers apply changes by running CLI commands against production.

    Explanation

    GitOps treats the repository as the single source of truth for desired state, with an automated agent continuously reconciling the environment to match. Changes are made by merging to Git rather than by manual commands or out-of-band edits.

  181. Question 181Automation, DevOps, and Infrastructure as Code

    Two pipeline jobs start terraform apply against the same remote backend at the same moment. Which mechanism prevents concurrent writes to the shared state?

    1. The checksum entries in .terraform.lock.hcl.
    2. Passing the -refresh=false flag to apply.
    3. Provider version pinning in required_providers.
    4. State locking provided by the backend. Correct answer.

    Explanation

    A remote backend that supports locking takes a lock for the duration of an operation so a second writer fails fast instead of corrupting state. Version pinning and the dependency lock file control which provider builds are used, and refresh behaviour does not serialize writers.

  182. Question 182Automation, DevOps, and Infrastructure as Code

    A production database is managed by Terraform, and the team must guarantee that no routine configuration change can delete it. Which control addresses this most directly?

    1. Adding a depends_on reference to the VPC.
    2. Adding lifecycle { prevent_destroy = true } to the resource. Correct answer.
    3. Marking the resource as sensitive.
    4. Moving the resource into a child module.

    Explanation

    The prevent_destroy lifecycle argument makes any plan that would destroy the resource fail, forcing deliberate human action before deletion. Dependencies, sensitivity marking and module structure affect ordering, output masking and organisation respectively.

  183. Question 183Automation, DevOps, and Infrastructure as Code

    An organisation needs fully isolated dev, staging and production environments with separate credentials and separate state. Which design is most appropriate?

    1. Separate root configurations with a separate state backend per environment. Correct answer.
    2. Terraform workspaces sharing one backend under the default workspace name.
    3. One workspace with per-environment variable files.
    4. A single state file using name prefixes for each environment.

    Explanation

    Separate root configurations backed by separate state locations give the strongest isolation of credentials, permissions and blast radius, and are the usual choice for production-grade layouts. Sharing a single state couples environments so a failed apply can affect all of them.

  184. Question 184Automation, DevOps, and Infrastructure as Code

    Which control best prevents an IaC change from creating a publicly readable storage bucket in production?

    1. Requiring two approvals on the pull request.
    2. Scheduling nightly terraform plan runs.
    3. A policy-as-code check that blocks plans creating public buckets. Correct answer.
    4. Running terraform validate in the pipeline.

    Explanation

    Policy-as-code evaluates the planned changes against explicit rules and fails the pipeline before the violating change is applied, which is deterministic rather than dependent on reviewer attention. Validate checks syntax, and nightly plans detect issues only after the fact.

  185. Question 185Cost Management and Governance

    Which statement best describes how cloud adoption changes an organization's cost structure?

    1. Depreciation schedules are applied to virtual machine instances.
    2. Long-term data center leases are converted into fixed annual license fees.
    3. Hardware refresh cycles are shortened to eighteen months.
    4. Upfront hardware purchases are replaced by recurring usage-based fees. Correct answer.

    Explanation

    Cloud computing converts upfront capital purchases into variable operating expense billed per use. This removes large depreciation cycles and ties spend directly to consumption.

  186. Question 186Cost Management and Governance

    What is the primary purpose of applying cost allocation tags to cloud resources?

    1. To encrypt billing data at rest.
    2. To reduce the per-unit price of compute.
    3. To enforce multi-factor authentication for billing access.
    4. To attribute charges to the owning team or project. Correct answer.

    Explanation

    Cost allocation tags attach business metadata to resources so billing data can be grouped by team, project, cost center, or environment. Without them, a shared account produces a single undifferentiated bill.

  187. Question 187Cost Management and Governance

    A team wants early warning before monthly cloud spend crosses an agreed ceiling. Which control fits that goal?

    1. A resource lock on the production subscription.
    2. A reserved instance purchase.
    3. A budget alert set below the spending limit. Correct answer.
    4. A service-level agreement with the provider.

    Explanation

    Budgets track planned spend and raise alerts at defined thresholds, providing notice before a ceiling is breached. They are a monitoring control and do not change prices or block deployments.

  188. Question 188Cost Management and Governance

    What does right-sizing a workload mean?

    1. Moving the workload to a region with lower latency.
    2. Purchasing a discount plan for the instance family.
    3. Increasing instance size to improve availability.
    4. Matching provisioned capacity to observed demand. Correct answer.

    Explanation

    Right-sizing means choosing capacity that matches measured demand, removing both over-provisioning waste and under-provisioning risk. It is a recurring optimization activity based on utilization data.

  189. Question 189Cost Management and Governance

    Which practice reports cloud consumption costs to teams for awareness without transferring budget responsibility?

    1. Anomaly detection
    2. Chargeback invoicing
    3. Showback reporting Correct answer.
    4. Consolidated billing

    Explanation

    Showback communicates consumption cost to teams so they can influence behavior, but no money moves between budgets. Chargeback is the variant that actually debits the consumer.

  190. Question 190Cost Management and Governance

    For which workload pattern is a one-year commitment discount economically appropriate?

    1. For steady-state workloads with predictable usage. Correct answer.
    2. For batch jobs that can tolerate interruption.
    3. For workloads that run only during a two-week campaign.
    4. For development environments that are stopped nightly.

    Explanation

    Commitment discounts trade a term obligation for a lower rate, so they pay off only when usage is sustained and predictable across the term. Spiky, interrupted, or short-lived workloads rarely consume enough to justify the commitment.

  191. Question 191Cost Management and Governance

    Which workload is the strongest candidate for interruptible capacity?

    1. A compliance archive with strict durability requirements.
    2. A checkpointed, fault-tolerant batch rendering job. Correct answer.
    3. A single-node domain controller.
    4. A transactional database serving customer orders.

    Explanation

    Interruptible capacity is discounted because the provider may reclaim it at short notice. It suits work that can checkpoint and resume, not stateful services with availability or durability guarantees.

  192. Question 192Cost Management and Governance

    A team moves 40 TB of analytics data each month from cloud storage to its on-premises data center. Which cost element must be modeled explicitly in the design?

    1. Storage durability charges.
    2. Ingress bandwidth charges.
    3. Data egress charges. Correct answer.
    4. Compute reservation charges.

    Explanation

    Outbound data transfer, known as egress, is billed separately from storage and compute and scales with the volume moved off the provider. Inbound transfer is typically free, so ingress is not the exposure here.

  193. Question 193Cost Management and Governance

    What is a primary financial benefit of consolidated billing across the accounts of one organization?

    1. Network traffic between regions becomes free of charge.
    2. Volume discounts and unused commitments are pooled. Correct answer.
    3. Identity policies are automatically replicated to all accounts.
    4. Each account receives a fully isolated invoice with no shared metadata.

    Explanation

    Consolidated billing rolls member accounts into one invoice so tiered volume thresholds and unused commitment capacity apply across the whole organization. Accounts remain distinct for identity and resource isolation.

  194. Question 194Cost Management and Governance

    An anomaly detector flags a 300 percent overnight rise in storage spend. What is the most appropriate first response?

    1. Disable all storage services in the account.
    2. Purchase additional savings plans to offset the increase.
    3. Raise the anomaly threshold so the alert stops firing.
    4. Check deployment and usage records to find the cause. Correct answer.

    Explanation

    An anomaly alert is a signal that requires triage against change and usage data before any spend or configuration action. Diagnosing the cause first avoids both masking the problem and disrupting live services.

  195. Question 195Cost Management and Governance

    Which metric best measures cost efficiency while transaction volume grows?

    1. Total monthly cloud spend.
    2. Count of active cost allocation tags.
    3. Cost per thousand transactions. Correct answer.
    4. Number of provisioned instances.

    Explanation

    Unit economics divide cost by a business output, so efficiency can be assessed even as volume rises. Total spend and resource counts increase with success and reveal nothing about efficiency.

  196. Question 196Cost Management and Governance

    Which activity belongs to the Inform phase of the FinOps lifecycle?

    1. Applying commitment discounts to steady workloads.
    2. Negotiating an enterprise discount agreement.
    3. Building dashboards that show teams their spend. Correct answer.
    4. Automating shutdown of idle development environments.

    Explanation

    The Inform phase covers visibility, allocation, and reporting so teams can see and understand their cost. Optimization actions belong to Optimize and rate negotiation sits with Operate.

  197. Question 197Cost Management and Governance

    Which governance mechanism prevents costly non-standard resources from being created at all?

    1. A weekly cost report emailed to finance.
    2. A guardrail that blocks unapproved resource types. Correct answer.
    3. A monthly budget alert triggered at 80% of forecasted spend.
    4. A tag dictionary maintained in a shared spreadsheet.

    Explanation

    Preventive guardrails deny the creation of unapproved or oversized resource types, so cost is avoided before it is incurred. Alerts, reports, and tag dictionaries are detective or informational controls that act after deployment.

  198. Question 198Cost Management and Governance

    An organization holds a commitment covering far more capacity than it uses, and unused commitment cannot be resold. What is the financial consequence?

    1. The unused portion is refunded at the end of the term.
    2. The commitment expands to cover additional services at no extra cost.
    3. The organization pays for the full commitment despite lower usage. Correct answer.
    4. The commitment automatically converts to on-demand rates for unused hours.

    Explanation

    A commitment is a fixed obligation, so unused capacity is still billed and the effective unit cost of what is actually consumed rises. Providers generally do not refund, flex, or convert unused commitment.

  199. Question 199Cost Management and Governance

    Which treatment makes a three-year on-premises versus cloud total cost of ownership comparison defensible?

    1. Comparing only the monthly compute price of the two options.
    2. Amortizing hardware and facility costs while including egress and labor. Correct answer.
    3. Excluding staff labor on the assumption that the provider performs all operational work.
    4. Charging the full hardware purchase to year one while counting cloud spend for one year.

    Explanation

    A defensible TCO compares like with like over the same period, spreading capital costs across their useful life and including every operating line such as transfer, support, and staff. Omitting cost lines or mismatching time horizons biases the comparison.

  200. Question 200Cost Management and Governance

    Shared platform costs such as logging, networking, and security tooling must be distributed to product teams. Which allocation approach is most defensible?

    1. Assigning all shared costs to the platform team's budget.
    2. Allocating shared costs by a documented usage driver. Correct answer.
    3. Allocating shared costs in proportion to each team's headcount.
    4. Splitting shared costs evenly across all teams regardless of consumption.

    Explanation

    Shared cost allocation holds up when it uses a documented, causally related driver that teams can influence and verify. Even splits, headcount, or leaving costs unallocated break the link between consumption and cost.

CCP-C flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

NIST cloud essential characteristics
NIST SP 800-145 lists five: on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service. These characteristics distinguish cloud computing from traditional hosted infrastructure.
IaaS vs PaaS vs SaaS
In IaaS the provider manages virtualization through networking while the customer manages the operating system upward; PaaS adds a provider-managed OS and runtime; SaaS delivers a complete provider-managed application.
Shared responsibility model
The provider is responsible for security of the cloud, covering facilities, hardware, and hypervisor. The customer is responsible for security in the cloud: data, identities, configuration, and patching above the service boundary.
Horizontal vs vertical scaling
Horizontal scaling adds or removes instances, called scale out and scale in, and is limited mainly by application design. Vertical scaling resizes a single instance and is capped by hardware limits.
High availability versus fault tolerance
High availability uses redundancy to mask failures, allowing brief interruption during failover. Fault tolerance keeps workloads running with no interruption through fully redundant, simultaneously active components.
Principle of least privilege
Grant each identity only the permissions needed for its task, for the minimum time required, and review entitlements regularly. This limits the blast radius if credentials are compromised.
Envelope encryption with KMS
A data key encrypts the data and a master key encrypts the data key; the wrapped data key is stored alongside the ciphertext, so the master key never leaves the key service.
Encryption at rest vs transit
At-rest encryption protects stored data using algorithms such as AES-256. In-transit encryption protects data moving across networks with TLS 1.2 or later, preventing interception and tampering.
Stateful versus stateless firewalls
Stateful security groups remember connection state and automatically allow return traffic. Stateless network ACLs evaluate every packet independently and require explicit inbound and outbound rules.
CIA triad
Confidentiality, integrity, and availability are the three core security objectives: preventing unauthorized disclosure, preventing unauthorized modification, and ensuring timely access to data and systems.
Blue-green deployment
Two identical environments run side by side; traffic switches from blue to green after the new version passes validation, and rollback is achieved by switching traffic back.
RPO versus RTO
Recovery point objective is the maximum acceptable data loss measured in time. Recovery time objective is the maximum acceptable service downtime. Both drive backup frequency and architecture choices.
Immutable infrastructure
Servers are never patched in place; instead a new image is built and deployed to replace the old one. This eliminates configuration drift and makes rollbacks predictable.
Canary release
A new version is exposed to a small percentage of live traffic first. Metrics are compared against the baseline and the rollout either expands gradually or is aborted automatically.
Configuration drift
Configuration drift occurs when live environment settings diverge from the declared source of truth, usually through manual changes. Drift detection compares actual state to desired state and flags differences.
CIDR block notation
An IPv4 CIDR suffix gives the number of network bits: /24 provides 256 addresses, /25 provides 128, /26 provides 64, and /27 provides 32, with usable hosts reduced by reserved addresses.
L4 versus L7 load balancing
Layer 4 balancers route on IP address and port without inspecting payloads, giving high throughput. Layer 7 balancers inspect HTTP headers and paths to support content-based routing and TLS termination.
NAT gateway purpose
A NAT gateway lets instances in private subnets initiate outbound internet traffic while blocking unsolicited inbound connections, because address translation is one-way unless a mapping exists.
Content delivery network caching
A CDN copies content to edge locations close to users, reducing latency and origin load. Cache TTL headers determine how long objects are served before revalidation against the origin.
Site-to-site VPN versus dedicated link
A VPN runs encrypted tunnels over the public internet with variable latency. A dedicated private circuit bypasses the internet, offering consistent latency and bandwidth at higher cost and longer lead time.
Block, file, and object storage
Block storage presents raw volumes to one host; file storage presents shared hierarchical directories over protocols such as NFS; object storage stores immutable blobs with metadata accessed by key over HTTP.
Durability versus availability
Durability is the probability that data is not lost over time, often expressed in nines. Availability is the proportion of time the service is reachable. Object stores typically exceed availability with durability.
RAID 5 versus RAID 6
RAID 5 tolerates one disk failure using distributed parity. RAID 6 tolerates two simultaneous failures with dual parity. Each sacrifices usable capacity equal to one or two drives respectively.
Storage lifecycle policy
Rules that automatically transition objects between hot, infrequent-access, and archive tiers and then expire them, based on object age, reducing cost as access frequency declines over time.
Synchronous versus asynchronous replication
Synchronous replication acknowledges a write only after the replica confirms it, giving near-zero data loss but higher latency. Asynchronous replication confirms locally first, giving lower latency but possible loss up to the RPO.
Declarative versus imperative IaC
Declarative tools such as Terraform and CloudFormation describe the desired end state and compute the required steps. Imperative tools execute explicit ordered commands such as scripts and ad hoc CLI calls.
Infrastructure as code state file
Terraform records managed resource identities in a state file so it can map real resources to configuration, compute diffs, and detect drift. The file must be stored securely and shared with locking.
Continuous integration
Developers merge code frequently into a shared branch where automated builds and tests run on every commit, surfacing integration defects early and keeping the mainline releasable.
Observability three pillars
Logs record discrete events, metrics are numeric time series aggregated over intervals, and traces follow a single request across services. Together they support diagnosis of distributed systems.
Capital versus operating expenditure
Capital expenditure is the upfront purchase of assets depreciated over time. Operating expenditure is pay-as-you-go consumption. Cloud shifts spending from capex toward opex and improves cash-flow flexibility.
Reserved versus spot pricing
Reserved commitments trade a one- or three-year term for a discount on steady workloads. Spot capacity offers the deepest discount but can be reclaimed with short notice, suiting interruptible jobs.
Showback versus chargeback
Showback reports cloud costs to teams without moving money. Chargeback actually bills each team's budget. Both depend on accurate tagging and account or project segmentation.

CCP-C study guide

Cloud Concepts and Architecture

Essential characteristics

Cloud is defined by five attributes: on-demand self-service, broad network access, resource pooling (multi-tenancy), rapid elasticity, and measured/metered service. The classic trap: a description of a ticket-driven, single-tenant virtualised data centre is not a private cloud, because self-service and pooling are missing. "Private cloud" and "on-premises" are not synonyms; a private cloud can even be hosted by a third party for one organisation's exclusive use.

Service models and the responsibility line

IaaS hands you virtual compute, storage and network (EC2, Azure VMs, Compute Engine). PaaS manages OS and runtime (App Service, App Engine, managed Kubernetes). SaaS delivers a finished application. FaaS/serverless functions are an event-driven slice of PaaS. Responsibility moves up the stack: in IaaS the customer patches the guest OS, configures firewalls and manages identities; in SaaS the provider runs the application while the customer still owns data, access grants and endpoints.

Deployment models and resilience

Public, private, hybrid (workloads span both, joined by secure connectivity) and community. Multi-tenancy is the economic engine of public cloud and the source of its isolation questions. Distinguish scalability (planned capacity growth) from elasticity (automatic response to demand), horizontal/scale-out from vertical/scale-up, high availability (survive a component failure, using regions and availability zones/availability sets) from disaster recovery (survive loss of a site), and remember an SLA typically promises service credits, not business continuity.

Rule of thumb: assign any control by asking "could the provider fix this without my involvement?" No OS to patch in your console → provider's job. Anything you configured in your own account (IAM, data, network rules, keys) → yours, whatever the service model.

Cloud Security, Risk, and Compliance

Core model

Security work starts from confidentiality, integrity and availability, applied through the shared responsibility model. In IaaS the provider secures facilities, hardware and hypervisor; the customer secures the guest OS, patching, network rules, identities and data. As you move to PaaS and SaaS, more controls become inherited from the provider - but never identity, access design or data classification.

Controls candidates must recognise

Identity and access management with least privilege, role-based assignment, federation/SSO and mandatory MFA for privileged accounts. Encryption in transit (TLS) and at rest, with key management services or HSMs; know the difference between provider-managed and customer-managed keys, since compliance regimes may demand customer key custody or BYOK/HYOK. Network and application controls: security groups, web application firewalls, DDoS protection. Detection and response: audit logs (CloudTrail, Activity Log, Cloud Audit Logs), centralised SIEM, CSPM for misconfiguration, CASB for SaaS shadow IT, vulnerability scanning and patch cadence. Incident response phases: prepare, detect, contain, eradicate, recover, lessons learned.

Risk and compliance

Distinguish threat (who/what could cause harm), vulnerability (the weakness), and risk (likelihood × impact). Treat risk by mitigating, transferring, avoiding or accepting. Frameworks you will see named: SOC 2, ISO 27001, PCI DSS, HIPAA, FedRAMP, GDPR, plus data residency/sovereignty rules.

Trap: "the provider is PCI DSS certified, so our workload is compliant." False - attestations cover the provider's layer only; your configuration, logging and access design must independently satisfy the standard. Similarly, encryption is a control, not a compliance certificate.

Worked example: a publicly readable object-storage bucket holds customer records. Threat = anonymous retrieval; vulnerability = no authentication plus no encryption; risk = high likelihood × high impact. Mitigation = block public access, enforce policy-as-code, enable default encryption, add alerting on policy changes, then accept the residual risk formally.

Deployment, Provisioning, and Operations

Provisioning vs configuration vs orchestration

Provisioning creates resources (a VM, a bucket, a load balancer). Configuration shapes them (packages, OS settings, application config). Orchestration sequences many resources into a running system. Questions often blur these; read for whether the task is creating or shaping.

Release strategies

Recreate/big-bang (simple, downtime). Rolling (replace instances in batches, partial exposure to a bad version). Blue-green (two identical environments, switch traffic, instant rollback, roughly double the cost). Canary (send a small slice of traffic to the new version, then progressively widen). Feature flags decouple deploy from release. Environment parity - staging resembling production - is what makes these strategies trustworthy.

Operating the estate

Monitoring collects metrics, logs and traces; observability is the ability to answer new questions from that telemetry without redeploying. Know SLI (measured indicator), SLO (internal target) and SLA (contractual promise with credits), plus error budgets, which let teams trade reliability against release speed. Patching cadence, runbooks, immutable infrastructure (replace, don't patch in place) and chaos engineering belong here too.

Recovery vocabulary

RTO = maximum tolerable downtime. RPO = maximum tolerable data loss, measured in time. Backup protects against deletion and corruption; replication propagates them, so replication is not backup. Snapshots are typically incremental and regional; cross-region copies are a separate configuration.

Rule of thumb: RPO sets backup frequency; RTO sets recovery method. Daily backups imply up to 24 hours of data loss no matter how fast you can rebuild.

Trap: swapped RTO/RPO in scenario questions, and choosing blue-green (cost) when a canary (risk control) is what the scenario describes.

Networking and Connectivity

Virtual network anatomy

A VPC/VNet is an isolated IP range defined in CIDR (for example 10.0.0.0/16), subdivided into subnets, each with its own route table. A subnet is public when its route table sends 0.0.0.0/0 to an internet gateway; it is private when the default route points to a NAT gateway (egress only, no inbound from the internet). Peering joins two networks privately, and it fails by design if the CIDR ranges overlap - a favourite exam trap. Hub-and-spoke topologies use a transit gateway or virtual WAN to avoid a mesh of peerings.

Filtering and distribution

Security groups are stateful, allow-only rules attached to network interfaces. Network ACLs are stateless, support allow and deny, and are evaluated in numbered order per subnet. Stateful means return traffic is automatically permitted; stateless means you must write both directions. Load balancers operate at layer 4 (TCP, fast, no content awareness) or layer 7 (HTTP, path/host routing, TLS termination, sticky sessions). DNS services offer routing policies - simple, weighted, latency, geolocation, failover - which are health-check driven. CDNs cache at the edge to cut latency and origin load.

Connecting sites

Site-to-site VPN runs encrypted tunnels over the public internet: fast to set up, cheap, variable latency. Dedicated private circuits (Direct Connect, ExpressRoute, Interconnect) give predictable latency and bandwidth at higher cost and with weeks of lead time; many designs use both, VPN as backup. Private endpoints / PrivateLink expose provider services inside your network without traversing the internet.

Rule of thumb: "does the subnet need to accept inbound connections from the internet?" Yes → internet gateway. No, but it needs outbound patching and API calls → NAT gateway.

Trap: watch for egress data charges and the assumption that VPN guarantees latency - only dedicated circuits do.

Data Management and Storage

The three storage shapes

Block storage presents raw volumes to one compute instance with low latency (EBS, Azure managed disks, persistent disks) - used for boot volumes and databases. File storage offers shared POSIX/SMB semantics to many clients concurrently (EFS, Azure Files, Filestore). Object storage exposes an HTTP API over a flat namespace with rich metadata and effectively unlimited scale (S3, Blob Storage, Cloud Storage) - right for backups, media, logs and static sites, wrong for a database's random I/O.

Durability, availability, tiers

Durability is the probability data survives; availability is the percentage of time it can be read. Object stores advertise eleven nines of durability while availability sits near four nines - do not conflate them. Storage classes/tiers (hot, infrequent access, archive) trade retrieval latency and minimum storage duration for price, and lifecycle policies move objects automatically. Replication options span single-zone, multi-zone and cross-region. Snapshots are usually incremental and stored regionally; restoring creates a new volume. Snapshots are not a substitute for cross-region DR.

Data platforms

Relational databases give ACID transactions, joins and strong consistency, scaling mostly vertically with read replicas. NoSQL families (key-value, document, wide-column, graph) trade joins for horizontal scale and often eventual consistency. A data warehouse is columnar and optimised for analytics; a data lake stores raw files with schema-on-read; a lakehouse blends both. Caching layers cut latency and database load.

Trap: choosing object storage for workloads needing POSIX or block semantics, treating archive tiers as instantly retrievable, and assuming eventual consistency will not affect read-after-write logic.

Rule of thumb: classify by access pattern - random block I/O → block; concurrent shared files → file; massive unstructured objects over HTTP → object.

Automation, DevOps, and Infrastructure as Code

DevOps fundamentals

DevOps is a culture and practice set summarised by CALMS - culture, automation, lean, measurement, sharing - expressed operationally through continuous integration (merge often, build and test automatically) and continuous delivery (every change is deployable). Automation exists to remove toil, shorten feedback loops and make change smaller and reversible.

Infrastructure as Code

IaC defines infrastructure in versioned files instead of console clicks. Declarative tools (Terraform, CloudFormation, ARM/Bicep, Pulumi) describe the desired end state and the tool computes the diff; imperative scripts describe steps and are not inherently idempotent. Idempotency means running the same definition twice leaves the same result - a defining property of declarative tooling. Tools maintain state mapping definitions to real resources; the state must live in a remote, locked backend because it contains sensitive values and concurrent runs corrupt it. Drift is the divergence between code and reality caused by out-of-band manual changes; drift detection surfaces it, but IaC does not prevent it automatically.

Pipelines and governance

A typical pipeline: commit → build → unit tests → static and security scans → deploy to staging → integration/acceptance tests → approval gate → production. GitOps takes this further by making Git the single source of truth, with an in-cluster agent continuously reconciling live state to the repository (pull-based). Policy as code (OPA, Azure Policy, SCPs) enforces guardrails automatically. Secrets belong in a secret manager, never in repositories - and remember they also leak into IaC state files.

Trap: statements like "IaC eliminates drift" (it doesn't; manual changes still happen) and confusing declarative convergence with imperative scripting.

Rule of thumb: if a task must be repeatable and reviewable, it belongs in code; if it must be identical every run, it must be idempotent.

Cost Management and Governance

The economics

Cloud converts capital expenditure (buying hardware up front) into operating expenditure (paying for consumption). Total cost of ownership must include migration, training, operations effort, licensing and - the classic omission - data egress and inter-region transfer. FinOps runs in three repeating phases: inform (visibility and allocation), optimise (right-size, discount, eliminate waste), operate (budgets, anomaly detection, continuous accountability).

Purchasing instruments

On-demand pricing is flexible and dearest. Committed-use discounts (reserved instances, savings plans, committed use) trade a one- or three-year commitment for a large discount; convertible/flexible variants give less discount for the freedom to change family or region. Spot/preemptible capacity offers the deepest discount but can be reclaimed at short notice, so it suits batch, CI and stateless work only. Hybrid benefit schemes let you bring existing licences. Know that reservations are usually a billing discount, not a capacity guarantee - capacity reservations are a separate product.

Waste and allocation

Right-sizing (matching instance size to measured utilisation), autoscaling, scheduling non-production shutdowns, lifecycle tiering, and deleting unattached disks, aged snapshots and idle load balancers are the recurring savings levers. Tagging (labels) drives cost allocation; unenforced tag policies produce untaggable spend, so enforce them at creation with policy. Distinguish showback (reporting costs to teams) from chargeback (actually billing them internally).

Governance guardrails

Organise accounts/subscriptions/projects under folders or an organisation, apply service control policies or Azure Policy to restrict regions, SKUs and services, set quotas and budgets with alerts, and consolidate billing to capture volume tiers centrally.

Trap: assuming a stopped instance costs nothing (attached storage still bills), or that spot capacity suits a stateful database.

Rule of thumb: steady, predictable load → commit; spiky or interruption-tolerant → spot; unknown → on-demand plus a budget alert.

CCP-C cheat sheet

Quick reference

Cloud Concepts and Architecture

  • NIST SP 800-145 defines five essential characteristics: on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service.

  • Service models stack responsibility: IaaS (provider handles virtualization through networking), PaaS (provider adds OS, runtime, middleware), SaaS (provider adds the application).

  • Four deployment models exist: public, private, hybrid, and community cloud.

  • Shared responsibility: the provider secures the cloud (facilities, hardware, hypervisor); the customer secures what runs in the cloud (data, identities, configuration, patching above the boundary).

  • Availability percentage to monthly downtime on a 30-day month: 99.9% = ~43.2 min, 99.95% = ~21.6 min, 99.99% = ~4.3 min.

  • Scale up/down = change the size of one instance; scale out/in = add or remove instances behind a load balancer.

  • High availability masks failure with redundancy and may allow brief interruption; fault tolerance continues with zero interruption via simultaneously active redundant components.

  • Region selection is driven by latency to users, data-residency compliance, service availability, and price; multi-region designs improve disaster recovery and reduce latency.

  • Elasticity matches capacity to demand automatically; scalability is the underlying ability to grow, which may be manual.

Quick reference

Cloud Security, Risk, and Compliance

  • CIA triad: confidentiality (no unauthorized disclosure), integrity (no unauthorized modification), availability (timely access).

  • Least privilege: grant only the permissions required for the task, for the shortest duration, and prefer temporary role credentials over long-lived access keys.

  • Envelope encryption: a data key encrypts the data; a master key in the key management service encrypts the data key; the wrapped data key is stored with the ciphertext.

  • Encryption at rest commonly uses AES-256; encryption in transit uses TLS 1.2 or later.

  • Stateful controls (security groups) remember connection state and auto-allow return traffic; stateless controls (network ACLs) need explicit inbound and outbound rules.

  • Common attestations to request from a provider: SOC 2 Type II, ISO/IEC 27001, PCI DSS, HIPAA BAA, and GDPR data processing terms.

  • The compliance boundary moves with the service model: the customer patches the guest OS in IaaS but not in SaaS.

  • Defense in depth layers identity, network, endpoint, application, and data controls so a single control failure does not expose the asset.

  • Data classification (public, internal, confidential, restricted) determines the encryption, access, and retention controls applied.

Quick reference

Deployment, Provisioning, and Operations

  • Provisioning creates infrastructure resources; configuration management sets software and settings on resources that already exist.

  • Blue/green swaps traffic between two identical environments; rollback is a traffic switch back to the previous environment.

  • Canary releases send a small percentage of traffic to the new version first, compare error and latency metrics to baseline, then expand or abort.

  • Rolling deployments replace instances in batches, keeping capacity available but running mixed versions during the rollout.

  • RPO = maximum tolerable data loss measured in time; RTO = maximum tolerable downtime. Both set backup frequency and recovery architecture.

  • Immutable infrastructure replaces servers with newly built images rather than patching in place, eliminating configuration drift.

  • MTTR is mean time to repair; MTBF is mean time between failures. Lower MTTR and higher MTBF both raise availability.

  • Autoscaling policies: target tracking holds a metric at a target value; step scaling adds capacity in tiers based on alarm magnitude.

  • A tested restore, not the backup job's success message, is the only proof that recovery meets RTO.

Quick reference

Networking and Connectivity

  • IPv4 CIDR host counts: /24 = 256 addresses, /25 = 128, /26 = 64, /27 = 32, /28 = 16; usable hosts are reduced by reserved network and broadcast addresses.

  • Public subnets route to an internet gateway; private subnets reach the internet outbound through a NAT gateway and accept no unsolicited inbound connections.

  • Security groups are stateful and instance-level; network ACLs are stateless and subnet-level, evaluated in numbered order with an implicit deny.

  • Layer 4 load balancers route on IP and port for throughput; layer 7 balancers inspect HTTP host, path, and headers for content-based routing and TLS termination.

  • Site-to-site VPN runs encrypted tunnels over the public internet with variable latency; a dedicated private circuit gives consistent latency and bandwidth at higher cost and longer lead time.

  • A CDN caches content at edge locations near users, cutting latency and origin load; cache TTL and invalidation rules govern freshness.

  • Common ports: 22 SSH, 3389 RDP, 80 HTTP, 443 HTTPS, 3306 MySQL, 5432 PostgreSQL.

  • VPC peering connects two networks directly and is non-transitive; a transit gateway or hub-and-spoke design is required to route many networks together.

  • DNS record types: A maps a name to IPv4, AAAA to IPv6, CNAME aliases one name to another, MX directs mail, and TTL controls resolver caching duration.

Quick reference

Data Management and Storage

  • Block storage presents raw volumes to a single host; file storage presents shared hierarchical directories (NFS/SMB); object storage stores immutable blobs with metadata addressed by key over HTTP.

  • Durability is the probability data is not lost over time; availability is the fraction of time the service is reachable. Object stores are typically far more durable than they are available.

  • Storage tiers trade retrieval latency and cost: hot, infrequent access, cold, and archive, with retrieval fees and minimum-duration charges in colder tiers.

  • Lifecycle policies automatically transition objects between tiers and expire them based on age, reducing cost as access frequency declines.

  • RAID 0 stripes with no redundancy; RAID 1 mirrors; RAID 5 tolerates one disk failure with distributed parity; RAID 6 tolerates two with dual parity; RAID 10 mirrors then stripes.

  • 3-2-1 backup rule: three copies of data, on two different media types, with one copy offsite.

  • Synchronous replication acknowledges a write only after the replica confirms (near-zero data loss, higher latency); asynchronous replication acknowledges locally first (lower latency, possible loss up to the RPO).

  • Strong consistency guarantees every read sees the latest write; eventual consistency allows stale reads for a period in exchange for lower latency and higher availability.

  • A data lake holds raw schema-on-read files; a data warehouse holds modeled schema-on-write tables optimized for SQL analytics.

Quick reference

Automation, DevOps, and Infrastructure as Code

  • Declarative IaC (Terraform, CloudFormation) states the desired end state and the tool computes actions; imperative approaches (scripts, some Ansible tasks) execute explicit ordered commands.

  • Idempotency means running the same automation twice produces the same end state without duplicating resources; it is the core requirement for safe re-runs.

  • Terraform keeps a state file mapping declared resources to real resource IDs; it must be stored remotely with locking and access control to prevent concurrent corruption.

  • Drift detection compares live resource configuration against the declared template and flags divergence caused by manual changes.

  • CI builds and tests every commit on a shared branch; CD automatically promotes artifacts that pass gates into staging and production.

  • Deployment strategies in pipelines: rolling (batched replacement), blue/green (parallel environment swap), and canary (percentage traffic ramp).

  • Observability rests on logs (discrete events), metrics (numeric time series), and traces (request paths across services).

  • Secrets belong in a managed secret store injected at runtime, never committed to source control or baked into images.

  • Test pyramid: many fast unit tests, fewer integration tests, and a small number of end-to-end tests to keep feedback loops short.

Quick reference

Cost Management and Governance

  • Cloud converts capital expenditure (upfront assets, depreciated) into operating expenditure (pay-as-you-go consumption).

  • TCO includes compute, storage, network egress, software licenses, facilities, labor, and downtime risk, not just the hourly instance price.

  • Purchasing options: on-demand (no commitment), reserved or committed-use (discount for a 1- or 3-year term on steady load), spot (deepest discount, interruptible).

  • Right-sizing matches instance family and size to measured utilization; over-provisioned instances are the most common source of waste.

  • Tagging by cost center, project, environment, and owner enables cost allocation; untagged resources cannot be charged back.

  • Budgets and alerts notify on forecast or actual spend thresholds; they do not cap spending by themselves.

  • Showback reports costs to teams without moving money; chargeback actually bills the team budget.

  • Network egress is frequently billed while ingress is free; cross-region and internet data transfer dominate unexpected bills.

  • Guardrails via policy as code and organizational policy constraints prevent costly or non-compliant resource types before deployment.

CCP-C key terms

Cloud Computing

Delivery of computing resources such as servers, storage, and applications over a network on demand, where a provider owns the physical infrastructure and charges customers according to usage.

Related: Elasticity, Virtualization, Multi-tenancy
Virtualization

Software that divides physical hardware into isolated virtual machines or environments, allowing several workloads to run independently on one server and improving utilization.

Related: Multi-tenancy, Container, IaaS
Multi-tenancy

An architecture where one shared pool of infrastructure serves many customers, with logical isolation ensuring each tenant's data and activity remain separate from others.

Related: Virtualization, Cloud Computing
Elasticity

The ability to automatically add or remove capacity as demand rises and falls, keeping resources matched to workload needs and to cost.

Related: Scalability, Autoscaling
Scalability

The ability of a system to handle increased load by adding resources, either by making instances larger (vertical) or by adding more instances (horizontal).

Related: Elasticity, Autoscaling, Load Balancing
IaaS

A cloud service model that supplies virtual machines, storage, and networks which the customer configures and manages, while the provider operates the physical data center.

Related: PaaS, SaaS, Virtualization
PaaS

A cloud service model offering a managed runtime and middleware where customers deploy their own code without administering servers, operating systems, or scaling infrastructure.

Related: IaaS, SaaS, Infrastructure as Code
SaaS

A cloud service model delivering a complete, provider-managed application over a network, which customers use through a browser or API without managing infrastructure.

Related: IaaS, PaaS
Hybrid Cloud

Two or more distinct cloud environments, typically private and public, kept separate but connected so applications and data can move between them.

Related: Cloud Computing, Virtualization
Region

A geographic area containing multiple isolated data centers, selected to place workloads near users and to meet latency, availability, and legal requirements.

Related: Replication, Hybrid Cloud
Shared Responsibility Model

The division of security duties in cloud computing: the provider secures the infrastructure, while the customer secures data, identities, configurations, and workloads placed in it.

Related: Identity and Access Management, Encryption at Rest
Identity and Access Management

Policies and technologies that verify who or what is making a request and decide which resources and actions that identity is allowed to use.

Related: Least Privilege, Multi-factor Authentication
Least Privilege

A security principle of granting each user, service, or process only the permissions required for its task, limiting damage from compromised credentials or mistakes.

Related: Identity and Access Management
Multi-factor Authentication

An authentication control requiring two or more independent proofs of identity, such as a password plus a code from a device, before access is granted.

Related: Identity and Access Management
Encryption at Rest

Converting stored data into ciphertext so that disks, backups, or snapshots cannot be read without the correct decryption key.

Related: Encryption in Transit
Encryption in Transit

Protecting data while it moves between systems by encrypting the connection, so intercepted traffic cannot be read or altered.

Related: Encryption at Rest
Provisioning

The process of creating and configuring resources such as servers, networks, and storage so they become ready for use, whether done manually or automatically.

Related: Orchestration, Infrastructure as Code, Autoscaling
Orchestration

Automated coordination of many resources and tasks to complete a workflow, such as deploying a multi-tier application, scaling a cluster, or recovering a service.

Related: Provisioning, Autoscaling
Autoscaling

Automatically adjusting the number or size of running resources in response to metrics such as CPU use or queue length, so capacity tracks demand.

Related: Elasticity, Scalability, Orchestration
Service Level Agreement

A contract that states expected service levels, such as availability and response time, and defines remedies or credits when those targets are missed.

Related: Monitoring and Observability
Monitoring and Observability

Collecting metrics, logs, and traces to understand system health, detect anomalies, and diagnose causes of failures across distributed applications.

Related: Service Level Agreement, Autoscaling
Virtual Private Cloud

A logically isolated network inside a provider's cloud where you define address ranges, subnets, routes, and gateways for your resources.

Related: Subnet, CIDR, Security Group
Subnet

A portion of a network's IP address range assigned to one availability zone, used to separate application tiers and control routing.

Related: Virtual Private Cloud, CIDR, Security Group
CIDR

A notation that expresses an IP address range as an address plus a prefix length, such as 10.0.0.0/16, used for planning cloud networks.

Related: Virtual Private Cloud, Subnet
Security Group

A stateful virtual firewall attached to cloud resources that allows or denies traffic based on protocol, port, and source or destination.

Related: Virtual Private Cloud, Subnet
Load Balancing

Distributing incoming requests across several servers to improve availability and responsiveness, and to let capacity grow or shrink without changing client configuration.

Related: Autoscaling, Scalability
Object Storage

A storage model holding data as objects with keys and metadata in a flat namespace, accessed through APIs and suited to unstructured files, backups, and static content.

Related: Block Storage, Lifecycle Policy, Replication
Block Storage

Storage presented as raw volumes attached to a server, which the operating system formats with a file system; suited to databases and low-latency random access.

Related: Object Storage, Snapshot
Snapshot

A point-in-time copy of a volume or database used to restore data after loss or corruption, usually stored incrementally and billed as consumed capacity.

Related: Block Storage, Replication
Replication

Copying data to another location, zone, or region, continuously or on a schedule, to improve durability and support recovery or local reads.

Related: Snapshot, Region
Lifecycle Policy

Rules that automatically move data between storage tiers or delete it after defined periods, controlling retention and cost without manual work.

Related: Object Storage
Infrastructure as Code

Defining infrastructure in versioned, machine-readable files that tools apply automatically, replacing manual console or script configuration with repeatable, reviewable changes.

Related: Configuration Management, CI/CD, Provisioning
CI/CD

Automated pipelines that build, test, and release code changes frequently, giving fast feedback and reducing the risk of each deployment.

Related: Infrastructure as Code, Configuration Management
Configuration Management

Tools and practices that bring servers and software to a defined state, installing packages and applying settings, and correcting configuration drift over time.

Related: Infrastructure as Code, CI/CD
Container

A lightweight, isolated package holding an application and its dependencies, sharing the host operating system kernel and starting far faster than a virtual machine.

Related: Virtualization, CI/CD
Operational Expenditure

Ongoing spending on services as they are consumed, which scales with usage and avoids large upfront purchases of hardware.

Related: Total Cost of Ownership, Reserved Capacity
Total Cost of Ownership

A complete estimate of the costs of an option over its life, including purchase, operations, staffing, migration, and exit, used to compare cloud and on-premises approaches.

Related: Operational Expenditure, Reserved Capacity
Reserved Capacity

A discount obtained by committing to a defined amount of usage for a fixed term, trading flexibility for a lower unit price.

Related: Total Cost of Ownership, Autoscaling
Tagging

Attaching key-value labels to resources so costs, owners, environments, and projects can be identified, reported, and governed at scale.

Related: Total Cost of Ownership

CCP-C study plan

8 weeks · About 7 hours per week

Week 1

Cloud Concepts and Architecture

  • Diagram the shared responsibility model, marking which controls shift to the customer for IaaS, PaaS, and SaaS

  • Compare vertical and horizontal scaling and match three workload patterns to the correct approach

  • Explain high availability, fault tolerance, and disaster recovery using RTO and RPO values you define yourself

  • List the essential characteristics of cloud computing and give one concrete provider example of each

Your checkpoint

A one-page architecture cheat sheet mapping 10 core terms (multi-tenancy, elasticity, measured service, region, availability zone, etc.) to a definition and a real-world example each.

Week 2

Cloud Concepts and Architecture & Cloud Security, Risk, and Compliance

  • Map each cloud service model to the customer-versus-provider control boundary for compute, network, storage, and identity

  • Distinguish identity federation, single sign-on, and multi-factor authentication, then sketch a login flow that uses all three

  • Classify ten common security controls as preventive, detective, or corrective

  • Define the confidentiality, integrity, and availability triad with one cloud-specific failure example per leg

Your checkpoint

A control matrix listing 15 security controls, each tagged by service model, control type (preventive/detective/corrective), and the CIA leg it protects.

Week 3

Cloud Security, Risk, and Compliance

  • Apply a four-step risk process (identify, assess, treat, monitor) to a workload that stores regulated data

  • Differentiate SOC 2, ISO 27001, PCI DSS, HIPAA, and GDPR by scope and by who they actually bind

  • Explain encryption at rest, in transit, and in use, and place key management correctly in each

  • Describe the incident response lifecycle and state where provider responsibility ends and yours begins

  • List the five most common cloud misconfigurations and the control that prevents each

Your checkpoint

A written risk register with 8 risks scored by likelihood and impact, each with a named treatment and the control that owns it.

Week 4

Deployment, Provisioning, and Operations

  • Contrast the six migration strategies and select one for three given workloads with a stated justification

  • Sequence a provisioning workflow from template to running resource, naming the approval and validation gates

  • Distinguish blue/green, canary, and rolling deployments and state when each is unsafe

  • Define patching, backup, and monitoring cadences for a production workload in a single table

Your checkpoint

A one-page runbook for deploying and rolling back a three-tier application, including the verification step at every stage.

Week 5

Networking and Connectivity & Data Management and Storage

  • Compare VPN, dedicated interconnect, and peering by latency, cost, and appropriate use case

  • Design a virtual network layout with public and private subnets and state which resources belong where

  • Differentiate block, file, and object storage and select the right type for four given workloads

  • Explain replication versus backup versus archival, stating the recovery-point implication of each

Your checkpoint

A network and storage diagram for a two-tier web app showing subnets, the ingress path, and the chosen storage class for each data type with a one-line rationale.

Week 6

Data Management and Storage & Automation, DevOps, and Infrastructure as Code

  • Define a data lifecycle policy (hot, warm, cold, delete) with retention periods for three data categories

  • Explain infrastructure-as-code idempotency and state management, and describe what breaks when state drifts

  • Compare CI/CD pipeline stages and identify which stage catches which class of defect

  • Write a rollback strategy for a failed infrastructure-as-code deployment

Your checkpoint

A roughly 20-line infrastructure template (real YAML or HCL, or faithful pseudocode) that provisions a network plus compute, with a comment explaining each parameter.

Week 7

Cost Management and Governance & Automation, DevOps, and Infrastructure as Code

  • Apply a tagging scheme to allocate spend across three teams and identify untagged spend

  • Distinguish committed-use, spot, and on-demand pricing by workload suitability

  • Identify three cost-optimization levers and the risk each introduces, such as spot interruption

  • Explain the governance mechanisms of budget alerts, quotas, policy as code, and approval workflows

Your checkpoint

A cost-optimization memo for a fictional workload: spend breakdown, three levers with estimated savings, and the governance guardrail attached to each.

Week 8

Cloud Concepts and Architecture & Cloud Security, Risk, and Compliance & Deployment, Provisioning, and Operations & Networking and Connectivity & Data Management and Storage & Automation, DevOps, and Infrastructure as Code & Cost Management and Governance

  • Retake full-length timed practice exams and log every objective you miss

  • Re-derive the 30 most-missed facts from memory on blank paper

  • Recite a two-minute explanation of each of the seven domains without notes

  • Re-test every item on your gap list and retire the ones you now get right

Your checkpoint

A final gap list of no more than 10 items, each with a specific fix and a re-test date before exam day.

Milestones

Week 2: Explain the shared responsibility boundary for any service model out loud in under two minutes with no notes, and have the 15-control matrix complete.

Week 4: Score 70% or better on a 40-question timed set covering Domains 1-3, and explain every missed item correctly without notes.

Week 6: Pass a 30-question timed set covering Domains 4-6 at 75% or better, and list the CI/CD pipeline stages in order unprompted.

Week 8: Score 80% or better on two full-length timed mocks taken on separate days, with no single domain below 70%.

CCP-C exam day

Your checklist

Government-issued photo ID

Bring your primary photo ID (driver's license or passport). Many testing centers require the name on your ID to match your registration exactly, and some require a second form of ID. Confirm the exact ID requirements, including whether a second form is needed, with the certification body or its testing vendor when you receive your scheduling confirmation.

Appointment confirmation / scheduling permit

Print or save your appointment confirmation with the test center address, date, time, and confirmation number. Some centers ask to see it; even where they don't, it prevents address mix-ups on the day.

Directions, parking, and a backup route

Scout the test center location beforehand. Know where to park, how long security check-in takes, and have a backup route or transit plan in case of traffic or road closures.

Arrival plan with a buffer

Plan to arrive well before your scheduled start time to absorb check-in and security screening. Late arrival can mean forfeiting your seat - confirm the center's late-arrival cutoff with the provider.

Knowledge of what is prohibited in the testing room

Electronics, watches, phones, notes, and most personal items are typically barred from the testing room; many centers provide a locker. Confirm the specific prohibited-items list and locker availability with the provider.

A light snack and water for before/after (stored as allowed)

Eat a normal meal beforehand. If the center permits items in a locker or waiting area, keep a snack and water there for after the exam. Confirm the center's policy on food and drink.

Comfortable layers and a plan for room temperature

Testing rooms are often cool. Dress in layers you can adjust; avoid clothing with excessive pockets or metal that can slow security screening.

Any pre-approved testing accommodations documentation

If you requested accommodations, confirm approval well in advance and bring any authorization documentation the provider instructs you to carry. Accommodation requests generally must be made before exam day.

A pacing plan for the question count and time limit

Before test day, calculate your approximate per-question time from the official item count and time limit, and decide your flag-and-return strategy. Confirm the exact item count and time limit from your candidate handbook.

Sleep and caffeine routine set in advance

Protect a normal night's sleep in the days before the exam and keep caffeine intake to your usual amount. Do not test a new energy drink or supplement on exam day.

A post-exam plan for score reporting

Know how and when results are released (often on-screen at the end and/or by email) and what you need to do next if you pass or must retest. Confirm the score reporting and retest policy with the provider.

Plan the day

72 hours before

Reconfirm your appointment date, time, and test center address using the provider's scheduling system. Resolve any discrepancy immediately rather than on the morning of the exam.

72 hours before

Do a final review pass focused on your weakest domains using your own notes and the official content outline. Stop learning new material; consolidate what you already know.

72 hours before

Shift your sleep schedule so you are waking at your exam-day time, and plan your meals and travel so the morning is routine rather than improvised.

24 hours before

Pack your ID, confirmation, and any allowed items in one bag and set it by the door. Lay out clothing in advance.

24 hours before

Do a short, light review (flashcards or a summary sheet) and then stop. Avoid full-length practice tests the day before; fatigue hurts more than extra practice helps.

24 hours before

Confirm your route, parking, and departure time, including a buffer for traffic. Check the weather forecast.

Morning of

Eat a normal breakfast, hydrate normally, and avoid excess caffeine. Arrive at the center with time to spare for check-in and security.

Morning of

Do not cram in the parking lot. If you want a warm-up, read a short summary of your own high-yield points for a few minutes only.

During the test

Read the full stem before looking at the options, and identify what is actually being asked. Watch for absolutes and for the 'best' versus 'first' distinction in priority questions.

During the test

Answer every question - there is generally no penalty for guessing, so never leave an item blank. Flag uncertain items and return to them rather than stalling.

During the test

Check your pace against the clock at natural intervals (for example, after each quarter of the item count). If you are behind, tighten up rather than rushing at the end.

During the test

Take a brief scheduled break if the center allows one, to reset focus - confirm the break policy and whether the clock runs during breaks before you sit down.

Rules to check

  • Follow the testing center's check-in process exactly: present the required ID, allow security screening, and store prohibited items as directed. If your ID does not match your registration name, you may be turned away - resolve name discrepancies with the provider before exam day.

  • Treat the exam as a confidential, nondisclosure-protected assessment. Do not record, photograph, copy, or discuss specific exam content with others; doing so can void your result and affect eligibility.

  • Keep all electronics, notes, and study materials out of the testing room. Even having a phone accessible during an authorized break can be treated as a violation under some providers' rules - confirm the policy with the provider.

  • Break availability and whether the exam clock continues to run during a break vary by vendor and are not guaranteed. Confirm the break policy for your appointment in advance so you can plan your pacing.

  • Scoring is typically criterion-referenced (pass/fail against a standard) rather than curved against other candidates, so your result does not depend on how others perform. Confirm the scoring model and pass standard in the official candidate handbook.

  • Answer every item. If the exam is scored on the number of correct responses with no penalty for incorrect answers, an educated guess is always better than a blank - confirm this in the candidate handbook.

  • If you need to reschedule, cancel, or request accommodations, do it within the provider's stated window; last-minute changes may forfeit your fee. Exact deadlines and fees are set by the provider - confirm them directly.

  • If a technical problem, distraction, or illness affects your test, notify the proctor immediately and ask for the incident to be documented. Post-exam appeals are usually much harder without a contemporaneous record.

CCP-C FAQ

How many CCP-C practice questions are available?

The practice set has 200 CCP-C questions. Each one includes the choices, the correct answer, and an explanation, and the set is free to read with no signup.

How many CCP-C flashcards are available?

The flashcard set has 32 CCP-C cards. Each card shows the term and its definition.

Which topics does the CCP-C study material cover?

The CCP-C study material has 7 sections, including Cloud Concepts and Architecture, Cloud Security, Risk, and Compliance, and Deployment, Provisioning, and Operations.

What format facts does the catalog list for the CCP-C exam?

The catalog lists these item types: multiple choice for the CCP-C exam.