We're not seeing periodic 503 errors anymore. Everything is back to normal.
Posted Feb 06, 2024 - 16:36 CET
Update
We are continuing to monitor for any further issues.
Posted Feb 06, 2024 - 15:31 CET
Monitoring
The issue affected a single of our clusters (on fr-par) hosting the serverless containers. The gateway in front of users containers was the culprit. We've made a fix and will continue monitoring, but so far we're not seeing 503 errors. This bug also affected the creation of new functions/containers, and this is now fixed too. Functions/containers in "error" status during the issue might need to be redeployed.
Posted Feb 06, 2024 - 15:24 CET
Investigating
Some users might receive intermittent 503 errors ("upstream connect error" or "no healthy upstream") when calling the endpoints of their serverless functions/containers in fr-par. We're currently investigating. For now, the issue is likely limited to some specific namespaces.
Posted Feb 06, 2024 - 14:39 CET
This incident affected: Elements - Products (Serverless Functions).
