Žygimantas A.

Could you please provide a small project to reproduce the issue?

Did you have any chance to make a repro project? I'm running the application in VS almost every full working day and I have never experienced the issue (although still...

That's weird. The exception data after the stack trace emits job id 2, while the detail at the beginning of exception complains about id 3.

Aha, suspected as much.. Well, really hope to get a repro at some point so I can investigate it further. I'm really unsure how exactly you manage the transactions.

There is a package now for Npgsql5. Same versioning and features, just a different name because of a different dependency - `Hangfire.PostgreSql.Npgsql5`. Simply switch to that one until you're able...

Please do. The primary reason why I rewrote connection management was to actually release them once done with the jobs, but I'm not sure how the dashboard works (yet). If...

[This](https://github.com/npgsql/npgsql/issues/3274#issuecomment-761814327) and linked issues seem to be related. We don't really do anything ourselves with the connection string/setup.

It would be interesting to see the connection strings (with sensitive details modified, of course, so all the config params are visible), how hangfire is configured (both adding services and...

I will try disabling pooling and checking how it goes that way. With pooling on (default) I am still only using the same amount of connections as previously (2 -...

Thank you for clarification. Indeed it makes sense then to have pooling off, although I've seen some posts that this might also be needed if any other "middle-man" is used,...