Connection enters .runningQuery state and never exits it.

Have adopted the library to do direct SQL queries. At first I used a single shared connection. It was fast, but after about 30 minutes of production running, I found it would enter a .runningQuery state and never exit, jamming the whole app.

Passing a timeout option did not help, though I suspect the timeout is for the PostgreSQL DB itself, not the library code.

I have since moved to creating a new connection on every request. This is much more costly, but it at least does work. However, I do see gradual memory rise over a day (from about 12MB to 48MB). It is usable, but not nice. I suspect this is the same problem: connections entering the .runningQuery state, and never exiting the state, meaning all the associated state is kept around.

I know I could use a pool to make things faster, but the bigger concern is just this jammed state for the connection. Has anyone else seen this? Unfortunately, it takes a while to show up, so it is rare-ish. Makes testing more difficult. Perhaps the connection itself should have some sort of timeout. It would at least error and then continue on properly.