Soru

Zorluk: Çok zorAutomating Continuous Deployment Pipelines and Release Strategies

An enterprise Cloud Architecture team is establishing an automated progressive release pipeline for a mission-critical microservice deployed to Google Kubernetes Engine (GKE). The release process must integrate Cloud Build, Artifact Registry, Cloud KMS, Binary Authorization, Google Cloud Deploy, and Cloud Monitoring to enforce zero-trust artifact security and automated metric-driven canary verification. Arrange the continuous deployment operational steps in the correct chronological execution sequence from artifact compilation to 100% production traffic cutover.

  1. 1Cloud Build compiles the container image, pushes it to Artifact Registry, and triggers an Attestor to sign the image digest using Cloud KMS, generating a Binary Authorization attestation.
  2. 2Cloud Build executes a CLI call to create a Google Cloud Deploy release resource, referencing the attested container digest, target delivery pipeline, and Skaffold manifest configuration.
  3. 3Google Cloud Deploy renders Kubernetes manifests and applies them to GKE, where the Binary Authorization admission controller verifies the signature before starting the baseline 10% canary traffic allocation.
  4. 4Automated Cloud Deploy verification jobs analyze latency and error rate SLIs collected by Cloud Monitoring; upon validation, Cloud Deploy promotes the release to advance canary traffic to 50%.
  5. 5Following successful completion of the final canary soak phase and metric validation checks, Cloud Deploy executes the full rollout, routing 100% of live traffic to the new target revision.

Cevap

The correct execution order is: 1) Cloud Build compiles the image, pushes to Artifact Registry, and creates a Binary Authorization attestation with Cloud KMS; 2) Cloud Build registers a Cloud Deploy release referencing the Skaffold configuration and attested image digest; 3) Cloud Deploy renders manifests to GKE, passing Binary Authorization admission control to establish initial 10% canary traffic; 4) Cloud Deploy verification jobs evaluate Cloud Monitoring SLIs to approve advancing traffic to 50%; 5) Cloud Deploy completes the rollout, directing 100% of live production traffic to the new release revision.
The automated deployment pipeline requires strict sequential dependency ordering: artifacts must be built and cryptographically signed in Cloud Build via Binary Authorization and Cloud KMS before a release can be registered in Cloud Deploy; manifest rendering and GKE admission control must validate signatures at pod creation before the initial canary phase receives traffic; real-time Cloud Monitoring metric verification must confirm health before Cloud Deploy advances traffic percentages; and finally, complete cutover (100% traffic) occurs after all canary gates pass.

Adım Adım Çözüm

1
Compile container image and produce cryptographic attestation
Attestation signature is verified and linked to the image digest stored in Artifact Registry.
Security governance requires binary provenance and attestation prior to release registration.
2
Instantiate Cloud Deploy release object
Cloud Deploy registers the release metadata and initiates pipeline rendering.
Cloud Deploy manages progressive delivery targets based on declared Skaffold rendering contexts.
3
Deploy baseline canary workload to GKE
GKE Binary Authorization admission controller validates attestation and permits pod creation with 10% traffic.
Cluster admission enforcement prevents unverified images from executing while exposing a limited canary population.
4
Perform automated metric-driven canary verification
Cloud Monitoring telemetric health indicators validate SLO compliance, authorizing phase traffic promotion.
Progressive delivery relies on empirical observability data to advance traffic allocation safely.
5
Complete production rollout
100% of live production traffic is directed to the fully verified release target.
Successful completion of all verification gates concludes the automated delivery lifecycle.

Anahtar Kavram

Continuous Deployment Pipeline Execution and Progressive Delivery Sequence with Cloud Deploy and Binary Authorization
Tahmini Süre:3m 0s
Bu soruyu puanla