too_many_connections too many connections for role “…” — 53300

Symptoms

A role hit its per-role connection limit set by ALTER ROLE … CONNECTION LIMIT. PostgreSQL refuses the new connection with SQLSTATE 53300 (too_many_connections).

• The role’s active connections reached its CONNECTION LIMIT.
• New connections for that role are rejected.
• Other roles may still connect normally.

What the server log shows

FATAL:  too many connections for role "app_user"

Why PostgreSQL raises this — what the manual says

As the ALTER ROLE reference (Parameters) explains:

The role has reached the maximum number of concurrent connections set by its CONNECTION LIMIT (ALTER ROLE name CONNECTION LIMIT n); raise the limit or wait for some of the role’s existing sessions to disconnect.

Each role can carry a per-role cap on concurrent sessions. When the role’s active connection count reaches that cap, PostgreSQL rejects further logins for that role with 53300, even if the cluster-wide limit has room.

Common causes

• The role’s CONNECTION LIMIT is set too low for the workload.
• Connection leaks holding sessions open.
• A surge in concurrent clients using that role.

Relevant GUC parameters

How to fix it

1. Raise the limit: ALTER ROLE app_user CONNECTION LIMIT 200;.
2. Use a connection pooler (PgBouncer) to bound and reuse connections.
3. Find and fix connection leaks in the application.

Diagnostic query

SELECT rolname, rolconnlimit FROM pg_roles WHERE rolname = 'app_user';

Compare the limit against current connections in pg_stat_activity for that role.

Related & next steps

• SQLSTATE 53300 — Too Many Connections (full reference)
• Message: too many connections for database
• Glossary: Connection pool

Reference: PostgreSQL 18 — ALTER ROLE.