Compute dashboard availability degraded

Partial outage Resolved · Down 13m

Affected: compute

Lifecycle

1. Reported 29 Oct, 17:04 UTC operator
2. Confirmed 29 Oct, 17:07 UTC operator
3. Investigating 29 Oct, 17:07 UTC operator
4. Resolved 29 Oct, 17:18 UTC operator
5. Closed not reached

Timeline

1. Resolved 29 Oct 2026, 17:18 UTC operator
   The compute homepage and sign-in page are responding normally again, with six frontend instances online. We have increased warm capacity and made autoscaling more responsive. We are monitoring recovery.
2. Confirmed 29 Oct 2026, 17:07 UTC operator
   Requests to the compute dashboard are returning HTTP 503 errors. Some public pages and the API health endpoint remain available.
3. Investigating 29 Oct 2026, 17:07 UTC operator
   We are investigating the frontend and preparing additional serving capacity. We will share another update after verifying recovery.
4. Reported 29 Oct 2026, 17:04 UTC operator
   We are investigating reports of errors accessing compute.pearlresearch.ai.