Fix "Too many connections" in MySQL (error 1040)

ERROR 1040 (HY000): Too many connections

Common causes

• Slow queries or lock waits keep connections busy while new requests keep arriving
• Persistent connections (PDO::ATTR_PERSISTENT, mysqli p:) or connection pools that never release
• Too many PHP-FPM workers compared with max_connections (default 151)
• A traffic spike, bot crawl or cron job storm opening many connections at once
• A long wait_timeout leaving thousands of idle Sleep connections
• A per-user cap set by max_user_connections on shared hosting (ERROR 1203)

How to fix it

1. See who holds the connections. Log in (MySQL keeps one extra slot for an admin with CONNECTION_ADMIN or SUPER) and run SHOW FULL PROCESSLIST; Group by user, host and Command to see whether they are Sleep or long Query states.
2. Kill stuck or idle sessions. Use KILL <id>; for runaway queries or long Sleep sessions. This frees slots immediately while you fix the cause.
3. Fix the slow queries. Enable the slow query log and add indexes or rewrite the queries that pile up. Shorter queries release connections faster than any setting change.
4. Lower idle timeouts. Set wait_timeout and interactive_timeout to 300-600 seconds so abandoned connections close. Disable persistent connections in PHP unless you have measured a benefit.
5. Match workers to max_connections. Keep pm.max_children across all PHP-FPM pools (plus cron and workers) below max_connections. If you have the RAM, raise max_connections to 300-500 in my.cnf and restart.
6. Add caching or a proxy. A page cache or object cache cuts database hits per request. For very high traffic, ProxySQL or MaxScale can pool and limit connections.

my.cnf ([mysqld] section)

[mysqld]
max_connections     = 300
wait_timeout        = 600
interactive_timeout = 600

-- check usage at runtime:
-- SHOW GLOBAL STATUS LIKE 'Max_used_connections';
-- SHOW VARIABLES LIKE 'max_connections';

How to stop it happening again

• Watch Max_used_connections and alert when it approaches max_connections
• Keep the slow query log on and review it after each release
• Size PHP-FPM and worker pools against the database limit, not just CPU
• Rate-limit aggressive bots and stagger cron jobs that hit the database