KCNA · Domain 4: Cloud Native Architecture · 12% of the exam
Task 4.2: Cloud Native Ecosystem and Principles
What makes a system cloud native: microservices, immutable infrastructure, declarative APIs, autoscaling, serverless, and the CNCF projects that fill each part of the stack.
Cloud native principles, microservices, serverless and the CNCF landscape
Lesson coming
Sample questions
Try each one before opening the answer. Every option is explained, with the CNCF documentation page that proves it.
Question 1 · choose 1
An online shop is one large application. The checkout code changes every day while the catalog rarely does, yet each checkout fix redeploys and retests everything, and the catalog must scale with checkout traffic. Which architecture addresses this?
ASplit it into microservices that are deployed and scaled independently
BGive the single application more CPU and memory on a bigger, faster node
CRun the application as a DaemonSet so every node has a copy
DPut the application behind an Ingress with path rules
Show the answer and why
ASplit it into microservices that are deployed and scaled independently
Correct
Microservices break an application into independent services, each focused on one function, so each can be released and scaled on its own.
BGive the single application more CPU and memory on a bigger, faster node
Incorrect
Vertical scaling adds capacity but keeps one deployable unit, so every change still ships everything.
CRun the application as a DaemonSet so every node has a copy
Incorrect
A DaemonSet places a copy on each node; the application is still one unit released as a whole.
DPut the application behind an Ingress with path rules
Incorrect
Path routing changes how requests arrive; the application still ships and scales as one unit.
Independent deployment and independent scaling of parts that change and grow at different rates is the core benefit of microservices.
A web service must be able to add or remove replicas at any moment, and Pods may be restarted or moved between nodes at any time. Where should each user's shopping cart be kept?
AIn each replica's memory, with sticky sessions on the load balancer
BIn an external data store that every replica can reach
CIn a file in the container's writable layer
DIn a ConfigMap that the replicas update as carts change
Show the answer and why
AIn each replica's memory, with sticky sessions on the load balancer
Incorrect
Sticky sessions send a user back to the same replica, but a cart in its memory is lost when that replica is removed, restarted or moved.
BIn an external data store that every replica can reach
Correct
Cloud native apps get scaled, restarted and moved, so state belongs in storage reachable from anywhere, such as a database.
CIn a file in the container's writable layer
Incorrect
That file disappears when the container restarts and is invisible to other replicas.
DIn a ConfigMap that the replicas update as carts change
Incorrect
A ConfigMap is an API object for non-confidential configuration. It is not designed to hold large chunks of data and is capped at 1 MiB; for more, the docs point to a volume or a separate database or file service.
Keeping instances stateless and the state in a shared store is what lets replicas be added, removed and moved freely.
Developers at a company spend days setting up environments, pipelines and monitoring for each new service. Leadership wants a dedicated team to give them shared, self-service tools and workflows. What is this practice?
AChaos engineering
BVertical scaling
CPlatform engineering
DDevSecOps
Show the answer and why
AChaos engineering
Incorrect
Chaos engineering tests resilience by injecting failures.
BVertical scaling
Incorrect
Vertical scaling adds resources to a single instance.
CPlatform engineering
Correct
Platform engineering provides tools and workflows that help developers work faster and avoid repetitive setup.
DDevSecOps
Incorrect
DevSecOps brings security into DevOps culture; it is not a team that provides self-service tools.
Platform teams build internal platforms so that application teams can serve themselves.
Two services share database tables, so every schema change in one breaks the other and both must be released together. Which architectural property is missing?
AVertical scalability of the database
BLoose coupling between the services
CService discovery by DNS name
DImmutable storage volumes
Show the answer and why
AVertical scalability of the database
Incorrect
Bigger instances do not remove the dependency between the services.
BLoose coupling between the services
Correct
In a loosely coupled architecture, components are built independently from one another.
CService discovery by DNS name
Incorrect
Finding services by name does not untangle their shared schema.
DImmutable storage volumes
Incorrect
Immutability of storage does not untangle shared schemas.
Tightly coupled components change together; loose coupling lets each evolve on its own.
A payment API is called again by clients when a response times out. The team wants a repeated identical request to have no additional effect. Which property should the operation have?
AStatelessness of the API
BHorizontal scalability
CSelf-healing of the Pods
DIdempotence of the operation
Show the answer and why
AStatelessness of the API
Incorrect
A stateless service can still charge twice for two identical requests.
BHorizontal scalability
Incorrect
More replicas do not prevent duplicate effects.
CSelf-healing of the Pods
Incorrect
Self-healing restarts or replaces failed parts; it does not make retries safe.
DIdempotence of the operation
Correct
An idempotent operation executed several times with the same parameters has no additional effect.
Idempotent operations make retries safe, which matters in distributed, failure-prone systems.
A company creates a team that applies software engineering to operations, builds systems that keep production applications running reliably, and focuses on code that is already running. What is this role?
ADevOps, with a focus on getting code to production
BQuality assurance
CSite Reliability Engineering (SRE)
DChaos engineering
Show the answer and why
ADevOps, with a focus on getting code to production
Incorrect
DevOps concentrates on delivering code to production; SRE focuses on how it runs there.
BQuality assurance
Incorrect
QA tests software before release rather than operating it.
CSite Reliability Engineering (SRE)
Correct
SRE applies software engineering to infrastructure and operations and ensures that code running in production works properly.
DChaos engineering
Incorrect
Chaos engineering is one practice such a team may use; it is not the role itself.
SRE and DevOps overlap, but SRE centers on running production reliably.