We checked the Postgres HA cluster in your Hackathon (NEW, 2026) project. The Patroni nodes in the production environment appear stable right now - postgres-2 holds the leader lock, and postgres-1 and postgres-3 are both following it as secondaries. Could you clarify which environment you're seeing the unhealthy status in (production or g42), and share the specific error you get when trying to connect? Also, are you connecting externally via the TCP proxy or from another service within the project? That will help us narrow down the issue.