We've seen some instances where the driver does not seem to be able to detect that its connections are in CLOSE_WAIT and given that we set db.SetMaxOpenConns(4), once we have 4 connections stuck in CLOSE_WAIT, we need to restart the app.

This seems to happen after a sudden influx of many requests.

Our setup has the app connecting to Postgres through HAProxy which is backed by pgpool, so it's possible that it might be related to that, but the driver should still be able to detect stale connections.