Google Cloud us-west1 outage hit about 30 services after fiber maintenance cut capacity

On August 20, 2026, Google Cloud's us-west1 region in Oregon suffered an outage of about two hours and 22 minutes, from 8:00 to 10:22 a.m. Pacific, with some backlogs clearing after that. About 30 products were affected, including Compute Engine, Cloud Storage, Google Kubernetes Engine, Cloud Run, Cloud SQL, BigQuery and Cloud IAM. Google's incident report says scheduled fiber optic maintenance unexpectedly reduced network capacity between data centers in the region, and automated rerouting failed to shift traffic, causing congestion that cascaded into core systems.

This is a reminder that a single region can fail even when no individual data center goes down. Customers saw timeouts, elevated errors and failures to create new resources, and Google recommended failing over to other regions where possible. Businesses that run everything in one region had no choice but to wait.

Ask your provider or internal team which cloud regions your critical applications depend on, including identity, storage and databases, not just the application servers. Confirm whether a documented, tested failover to another region exists, how long it takes and who decides to trigger it. If a second region is too costly, at least confirm you have backups stored outside the primary region and a clear communications plan for customers during an outage.

Google says it will add stricter maintenance safety checks, circuit redundancy requirements and better automated regional failover. Those are fixes on Google's side; the decision to design for regional failure still rests with each customer.