Cloud Run "ContainerImageImportFailed" persistent since 2h - all deploys blocked, region europe-west1

Google Cloud Serverless Applications
1

Since 2026-09-02 ~13:53 UTC, every deploy to a Cloud Run service
in europe-west1 fails with ContainerImageImportFailed - Container import failed. No self-recovery after 2+ hours.

Environment

  • Region: europe-west1
  • Service: single specific service (project: multiple Cloud Run services, only one tested affected)
  • Deploy method tried: both gcloud run deploy --source . and --image=<digest>
  • Image built by Cloud Build with standard Dockerfile (Python 3.12-slim + postgres-client-18)

What I verified (evidence)

  1. Cloud Build succeeds - image is pushed correctly to Artifact Registry
  2. Image manifest valid: Docker v2, linux/amd64, 9 layers, ~230MB
  3. All layer blobs + config blob present in AR (HTTP 200 via /v2/.../blobs/<digest> with auth token)
  4. Same digest that imported OK 3h earlier now fails — reproducible on demand
  5. Redeploying older known-good images (from same day, before the outage started) also fails now
  6. YAML spec diff between last successful revision and failing rev shows only the image digest
    changed
  7. IAM: service-<projectNumber>@serverless-robot-prod.iam.gserviceaccount.com and
    @gcp-sa-artifactregistry both have serviceAgent role
  8. Google Cloud Status: no incident reported for the region/service

What I ruled out

  • Non-UTF8 filenames in source (checked, clean)
  • Foreign layers (native linux/amd64, no Windows base)
  • Artifact Registry cleanup policy (repo still 79GB, policy applied 24h ago but never executed, zero
    Delete events in audit log last 6h)
  • Corruption of a specific image (multiple different digests all fail identically)
  • Source code issue (reset to last known-good commit + redeploy still fails)

Impact

Service is live at 100% traffic on the previous revision (no user downtime),
but ALL new deploys are blocked. Cannot roll out any code updates.

Questions

  1. Anyone else seeing ContainerImageImportFailed on Cloud Run in europe-west1 today (2026-09-02)?
  2. Known issue with Cloud Run image importer this week?
  3. Any workaround besides opening a paid Standard Support case?

I can share revision names, image digests, YAML diffs, and HTTP HEAD
results on all blobs on request.