High connection latency via shared Pooler in us-west-1

This incident has been resolved.

The team noticed that some connection pools had workers stuck as a consequence of the previous issue. This could cause query failures. Stuck workers were now restarted.

We've removed the problematic cluster node and latency returned to the normal level. We are now monitoring.

We’re investigating high latency in us-west-1 affecting some connections to databases via our shared connection pooler.