HiveDeployReliability ledger
RELIABILITY (java × gcp)2026-07-10T12:45Zgate version v1

01_java_cloudrun#4

FAILED
bucket:AGENTclass:cloudrun-healthz-probe-contract

evidence:

Attributed 'network blip' at campaign close; instrumented re-runs then proved a real platform defect — the post-apply health probe required 200 on /healthz, but Google's front end reserves /healthz externally (404 to outside callers even when the container serves it). Affected every Cloud Run deploy. The only AGENT-attributed failure in the 64-run campaign.

  1. date2026-07-10T12:45Z
  2. deploy statusfailed (2113s)
  3. gate––––healthz=— connection=— logs=— metrics=—
  4. teardown cleanTrue

resolution:

Probe fallback fix (92c5518d) + java_app_cloudrun_gcp frozen module chain; class re-run green on deploy-health basis (scale-to-zero idle window makes the recency-scoped log stream empty by design — disclosed).

Recorded reason (scrubbed)

{"runId": "run-e67151c6", "message": "Post-apply health check FAILED: the deployed java service did not come up on its public endpoint after a successful terraform apply \u2014 the provisioning bootstrap (userdata/init) did not start the service or b