I have a self-hosted on VPC under AWS deployed quite a while ago. No issues were found until recently I notice the workflow worker keeps rebooting. Looking into the log file and notice the following errors:
The postgres_ssl_rejection_unauthorized has always been there and we will address it soon. But the DNS error is something new. How does Retool resolve the the code-executor and workflows-backend? I thought this will resolve back to localhost but seems somehow it tries to resolve to an IP and I can't locate the logic inside the docker configuration file.
I just notice in the docker compose ps I used to see temporal as a container item. But in the recent version where temporal is move to a separate yaml file, it seems I no longer see this temporal container entry. I check the compose.yaml and it is included. Is this correct behavior?
Hey @yikl - thanks for reaching out. Did you upgrade versions recently? I would start by double checking the value of the WORKFLOW_BACKEND_HOST environment variable and verifying that the two services are on the same network.
It might be helpful to see the contents of your compose file, regardless! If you want to do some additional digging, you can exec into the workflow worker and manually test DNS resolution.
