Connection Error in the logs of supavisor

[jahnavisana2812]

I've set up Supavisor on my PostgreSQL cluster hosted on Kubernetes. The connection pool is functioning properly, and I can make connections. However, I keep seeing this error in the Supavisor logs:

06:26:32.388 project=tenant user=user region=local mode=transaction type=single [warning] ClientHandler: socket closed with reason {:shutdown, :terminate_received}

Can you help me understand why I keep seeing this error? Is there any misconfiguration on my part, and is there a way to stop this error from appearing?

[jumski]

Hey @jahnavisana2812, I'm not a supavisor developer but working on problems with abrupt connection closures and pool saturation in my app.

Maybe my findings would be helpful for you:

{:shutdown, :terminate_received}

Usually denotes that the process has received a termination command and is following OTP’s structured, albeit sometimes hurried, shutdown sequence. While ideal in theory, this may still fall short in environments where resources must be reclaimed immediately.

{:shutdown, :socket_closed}

Typically signals that a socket has been fully shut, usually implying a hard close at the TCP level. This can inadvertently leave connection slots reserved if the process termination does not complete in a controlled manner.