Database became unhealthy even though CPU, memory, and disk usage looked normal
🔢DB Connections
auth
authMy Supabase database suddenly became unhealthy, and I’m trying to understand what may have caused it. When I checked the metrics around the time of the incident, I noticed that: CPU usage was around 60%; Memory usage was low; Disk usage was also low.
From the infrastructure metrics alone, nothing seemed critically overloaded, so I’m confused about why the database was marked as unhealthy. I also reviewed the logs, but they do not clearly explain the root cause. I saw some messages related to slow queries, broken pipe, and schema introspection, but I’m not sure whether these were the cause of the unhealthy state or just side effects.
I would like to better understand: What conditions can make a Supabase Postgres instance become unhealthy even when CPU, memory, and disk appear to be within normal levels?
Could this be related to:
- connection saturation,
- long-running queries,
- lock contention,
- internal Supabase services,
- PostgREST/schema introspection issues,
- or something else not visible in the basic resource charts?
Is there any way to identify the exact reason why the instance was marked unhealthy? Are there additional logs, metrics, or dashboards I should check to diagnose this more precisely?
Any guidance would be appreciated. I’m trying to understand whether this was caused by workload inside Postgres, an internal service issue, or a platform-level health check failure that is not obvious from the standard metrics.
Thanks in advance.
From the infrastructure metrics alone, nothing seemed critically overloaded, so I’m confused about why the database was marked as unhealthy. I also reviewed the logs, but they do not clearly explain the root cause. I saw some messages related to slow queries, broken pipe, and schema introspection, but I’m not sure whether these were the cause of the unhealthy state or just side effects.
I would like to better understand: What conditions can make a Supabase Postgres instance become unhealthy even when CPU, memory, and disk appear to be within normal levels?
Could this be related to:
- connection saturation,
- long-running queries,
- lock contention,
- internal Supabase services,
- PostgREST/schema introspection issues,
- or something else not visible in the basic resource charts?
Is there any way to identify the exact reason why the instance was marked unhealthy? Are there additional logs, metrics, or dashboards I should check to diagnose this more precisely?
Any guidance would be appreciated. I’m trying to understand whether this was caused by workload inside Postgres, an internal service issue, or a platform-level health check failure that is not obvious from the standard metrics.
Thanks in advance.
