Version: v2.33.5 81139b5
Hi Coder team,
We have deployed the Coder platform on an Azure Kubernetes Service (AKS) cluster, with PostgreSQL running as a pod and being used as the backend database.
The whole setup was working fine for 3 months and recently, we have been experiencing an issue where the Coder pod is repeatedly restarting. After checking the pod logs, it appears that the restarts are being triggered by an internal error.
we have attached a few relevant log snippets below that show the errors we are seeing. The complete logs are quite lengthy, so we have attached them separately as files for further investigation.
Could you please help us understand:
- What could be causing this internal error?
- Why is the Coder pod restarting as a result?
- Is this related to the PostgreSQL backend, Coder configuration, or another component?
- What steps can we take to troubleshoot and resolve the issue?
Any guidance on the root cause and recommended remediation would be greatly appreciated.
Thank you!
2026-09-08 01:19:09.357 [erro] pubsub: pubsub disconnected from postgres error=EOF
2026-09-08 01:19:09.383 [erro] pubsub: pubsub failed to dial postgres network=tcp address=postgresql.coder-nonprod.svc.cluster.local:5432 timeout_ms=0
2026-09-08 01:19:09.383 [erro] pubsub: pubsub failed to connect to postgres error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:09.530 [erro] coderd.chatd.processor: failed to acquire chats error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:10.413 [erro] pubsub: pubsub failed to dial postgres network=tcp address=postgresql.coder-nonprod.svc.cluster.local:5432 timeout_ms=0
2026-09-08 01:19:10.413 [erro] pubsub: pubsub failed to connect to postgres error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:10.504 [erro] coderd.chatd.processor: failed to acquire chats error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:11.511 [erro] coderd.chatd.processor: failed to acquire chats error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:11.511 [warn] coderd.inmem-provisionerd-coder-649cdcc84d-8tnph-1: heartbeat failed error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:11.511 [warn] coderd.gitsync: acquire stale chat diff statuses error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
2026-09-08 01:19:11.511 [warn] coderd: run replica update loop ...
error= get replicas:
github.com/coder/coder/v2/enterprise/replicasync.(*Manager).syncReplicas
/home/runner/work/coder/coder/enterprise/replicasync/replicasync.go:251
' - dial tcp XX.XX.XX.XX:5432: connect: connection refused
2026-09-08 01:19:11.605 [warn] coderd.inmem-provisionerd-coder-649cdcc84d-8tnph-0: heartbeat failed error="dial tcp XX.XX.XX.XX:5432: connect: connection refused"
Error_1.log
Error_2.log
Version:
v2.33.5 81139b5Hi Coder team,
We have deployed the Coder platform on an Azure Kubernetes Service (AKS) cluster, with PostgreSQL running as a pod and being used as the backend database.
The whole setup was working fine for 3 months and recently, we have been experiencing an issue where the Coder pod is repeatedly restarting. After checking the pod logs, it appears that the restarts are being triggered by an internal error.
we have attached a few relevant log snippets below that show the errors we are seeing. The complete logs are quite lengthy, so we have attached them separately as files for further investigation.
Could you please help us understand:
Any guidance on the root cause and recommended remediation would be greatly appreciated.
Thank you!
Error_1.log
Error_2.log