How to reset database cluster

Heya,

First, I suggest you find the source of those connections. Are they connections that were not closed properly, or do you have a lot of connections to your application?

If it’s the first, until you fix the problem, these connections will continue to appear and stay open. Only after that is done, you can be considered looking at the connections.

If your application has a lot of traffic, thus creating a lot of DB connections, then you’ll need to consider upgrading.

As for the connections they should clear automatically once they timeout.

If you’re hitting the max_connections limit often, and can’t restart the cluster directly from the dashboard (which DigitalOcean doesn’t support for managed databases), one workaround is to temporarily scale the cluster up and back down. This can help force a reset on lingering connections. I’ve done this once on a staging cluster where a buggy script flooded connections, a quick scale toggle cleared it without data loss. But yeah, the root cause is usually app-side connection handling.