too_many_connections remaining connection slots are reserved for non-replication superuser connections — 53300

Symptoms

The server is near its connection limit and the only free slots are reserved for superusers (via superuser_reserved_connections). A normal role’s connection is refused with SQLSTATE 53300 (too_many_connections).

• Appears when connections approach max_connections.
• Superusers can still connect; ordinary roles cannot.
• Usually a symptom of connection leakage or missing pooling.

What the server log shows

FATAL:  remaining connection slots are reserved for non-replication superuser connections

Why PostgreSQL raises this — what the manual says

Section 19.3.1 Connection Settings (superuser_reserved_connections):

“Whenever the number of active concurrent connections is at least max_connections minus superuser_reserved_connections, new connections will be accepted only for superusers.”

PostgreSQL keeps superuser_reserved_connections slots out of max_connections for superusers so admins can always get in. Once non-reserved slots are exhausted, ordinary connections are refused with 53300.

Common causes

• More client connections than the server is sized for (no pooling).
• Connection leaks in the application that never release sessions.
• A spike in traffic or a stuck pool holding connections open.

Relevant GUC parameters

How to fix it

1. Introduce a connection pooler (PgBouncer) so apps share fewer server connections.
2. Find and fix connection leaks; ensure pools have sane max sizes.
3. Raise max_connections only if the server has the memory headroom (requires restart).

Diagnostic query

-- Who is using the connections right now?
SELECT usename, state, count(*)
FROM pg_stat_activity
GROUP BY usename, state
ORDER BY count(*) DESC;

Related & next steps

• SQLSTATE 53300 — Too Many Connections (full reference)
• Message: sorry, too many clients already
• Glossary: Connection pooling

Reference: PostgreSQL 18 Section 20.3 “Connections and Authentication”.