A pool size of 10 for your production instances is pretty small, I’d probably double that. The error you’re seeing is relatively simple. You have a pool of 10 within your application, and a process had to wait over 1 second to even get a connection, which meant all 10 were busy for that full second. This tells me two things:
- You probably have too few connections and
- you may be holding on to connections too long, I’d check your db query metrics.
