PostgREST stuck Unhealthy after restart, DB confirmed healthy — pause/unpause helped others, hasn't fixed mine

[suchana-abstrabit]

Project ref: utptdqcimanwflgthgva

PostgREST has been showing "Unhealthy" on the project health dashboard
for over an hour, while Database, Auth, Realtime, Storage, and Edge
Functions all show "Healthy."

I restarted the project once via Project Settings > Infrastructure —
no change. Well past the stated "up to 5 minutes" recovery window for
recently restored projects.

Confirmed via Logs Explorer that PostgREST's own readiness endpoint is
returning 503:

HEAD | 503 | https://utptdqcimanwflgthgva.supabase.co/rest-admin/v1/ready
Timestamp: 2026-08-14T08:44:51.380Z

I also confirmed the underlying database itself is fully responsive —
ran a pg_stat_activity check via the SQL Editor and found no stuck or
long-running transactions, no lock waits. So this doesn't appear to be
a schema lock or resource exhaustion issue on the DB side — it looks
isolated to PostgREST specifically.

This is blocking all RPC calls (supabase.rpc(...)) from my backend with
503 upstream connect error responses.

I've seen a similar case (Discussion #42817) where a full pause +
unpause cycle (not just "Restart Project") resolved the same symptom.

I've also filed a support ticket (SU-446153) for this, posting here in
case anyone's hit the same thing recently or has additional troubleshooting
steps.

[k1ng-arthur]

If you are in us-east-2 then you may be affected by this https://discord.com/channels/839993398554656828/1409892722474614814/1537760632277835797 which the team is resolving