It's a mystery. We have two instances that serve a lot of traffic without problems, but sometimes the database connections spike, the instances stop serving traffic and I have no idea what causes it.

What could I do to investigate? Wading through the logs seems difficult without an idea of what to look for.

We tried with v0.7.x and 0.8.x and with and without docker.

Could be something related to the database? It's a postgresql