Elevated API latency in EU West

Updates

1. Resolved

   Oct 8, 10:00 UTC

   Latency has been normal for 20 minutes. This incident is resolved.

2. Monitoring

   Oct 8, 09:38 UTC

   The rollback is done and p95 latency is back under 200 ms. We'll keep watching for the next 20 minutes.

3. Identified

   Oct 8, 09:23 UTC

   A configuration change on the EU West load balancers sent all traffic through a single availability zone. We're rolling it back.

4. Investigating

   Oct 8, 09:12 UTC

   We're seeing p95 response times above 900 ms for API requests served from EU West. Other regions aren't affected. We're looking into it.