- Sources: incident, status page, discussion
- Summary: Google opened incident J5ia5t9p3g9Q5Wi7r8Ev with a begin time of 2026-09-01 14:44 UTC, scoped to a portion of the us-central1-b zone with the remaining zones in the region unaffected, and states that routine network fabric path maintenance triggered unexpected issues in two clusters, that the impacted network capacity has been recovered, and that maintenance in the region has been halted while proactive audits are carried out. Customer symptoms are stated as elevated packet loss and errors, and at the 18:36 UTC update Cloud Run and Google App Engine were still listed as impacted with errors described as lowered to pre-incident levels, while BigQuery, Compute Engine, Google Kubernetes Engine, Cloud SQL, Cloud Spanner, Cloud Bigtable, Cloud Dataflow, Virtual Private Cloud, Cloud NAT, Cloud Interconnect, Cloud Filestore, Apigee, Looker, Cloud Firestore and Google SecOps SOAR are listed as recovered. Google states that customers with service layer retries may see successful completions because only a portion of the zone is affected, and this is the second Google Cloud incident in under two weeks whose stated cause is planned maintenance removing network capacity, after the us-west1 incident of 2026-08-20 whose published report attributes it to scheduled fiber optic maintenance combined with automated rerouting that failed to redistribute traffic.
- Why it matters: The stated cause is a planned change window rather than a hardware fault, and it is the second such window in two weeks to remove network capacity in a Google Cloud region.
- Follow-up: Track whether Google publishes a post-incident report for J5ia5t9p3g9Q5Wi7r8Ev and whether the halted regional maintenance resumes.
send feedback on this story