Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team manages a stateless microservice running on Google Kubernetes Engine (GKE). During startup, each Pod takes 45 seconds to load an in-memory cache before it can handle requests, and under heavy load, containers occasionally enter a deadlock state. You must update the service while meeting these requirements:
Which deployment and health-monitoring configuration should you implement?
RollingUpdate strategy with maxSurge: 25% to hold 25% of traffic on the new version, and configure a liveness probe to remove unready Pods from the load balancer until the cache loads.A Kubernetes RollingUpdate strategy incrementally replaces old Pods with new Pods (allowing up to maxSurge: 25% extra Pods above the desired replica count during the update), paired here with a liveness probe intended to manage load balancer registration.
RollingUpdate continuously replaces old Pods with new ones as soon as new Pods pass readiness checks until 100% of replicas run the new version. Setting maxSurge: 25% only controls how many extra Pods can be created simultaneously; it does not implement a canary deployment that holds 25% of traffic on the new version for validation.RollingUpdate behavior: Designed for continuous, automated Pod replacement rather than staged, percentage-based canary promotion.A RollingUpdate with maxSurge: 25% would be suitable if you wanted an automated, zero-downtime replacement of all Pods without holding at a 25% canary stage—and only if paired with a proper readiness probe for cache warm-up.
standard strategy targeting GKE to deploy the new Pods, add a liveness probe that waits for the cache to finish loading, and omit the readiness probe so Pods register with the load balancer immediately.The Cloud Deploy standard deployment strategy replaces the existing application version directly without progressive traffic splitting, and this configuration relies solely on a liveness probe while omitting a readiness probe.
standard strategy performs an all-at-once deployment to the target runtime rather than splitting traffic between old and new versions, making it impossible to test the release with 25% of traffic first.SIGKILL / Pod restart) on failure rather than temporarily withholding traffic. If the 45-second startup exceeds the liveness probe's initial delay and failure threshold, Kubernetes restarts the Pod repeatedly in a CrashLoopBackOff.This approach would only fit a non-critical workload where immediate 100% rollout is acceptable, startup is instantaneous with no cache warm-up, and no progressive traffic testing is required.
canary strategy targeting GKE with a 25% initial phase before stable, add a readiness probe that returns HTTP 200 only after the cache loads, and add an independent liveness probe to detect deadlocks.A Cloud Deploy canary strategy combined with properly scoped Kubernetes readiness and liveness probes enables progressive traffic shifting on GKE while protecting request availability and container health during transitions.
canary strategy (using GKE service networking or Gateway API) creates an initial canary-25 phase that routes 25% of traffic to the new version while keeping the existing version serving the remaining 75%. After validation, you advance the rollout to the stable (100%) phase.200 on the readiness endpoint only after the 45-second cache finishes loading, new Pods remain excluded from load balancing until they can serve traffic.stable.Unlike a standard deployment or an unmanaged Kubernetes RollingUpdate, a Cloud Deploy canary pipeline explicitly pauses at the 25% mark for validation, while pairing distinct readiness and liveness probes ensures each probe handles its exact Kubernetes responsibility.
Recreate with a Pod Disruption Budget (PDB) reserving 25% capacity, and configure a readiness probe that checks downstream dependencies to confirm the Pod is healthy.The Recreate deployment strategy terminates all running Pods of the previous version before creating Pods for the new version, while a Pod Disruption Budget (PDB) defines minimum availability thresholds during voluntary cluster disruptions.
Recreate scales down all existing Pods to zero before starting new Pods. Because new Pods also need 45 seconds to load the cache, the service experiences a complete outage during every update.A Recreate strategy and PDB would be appropriate only if the application could tolerate downtime during updates, required strict single-version exclusivity (for example, exclusive file locks), and needed protection against node-drain evictions.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated