MySQL at peak runs out of connections with ERROR 1040. Replica lag is the quieter failure: it serves stale reads and logs nothing, so you need an alert for it. Start with Fix MySQL ERROR 1040 Too Many Connections.
- Monitor MySQL Replication Lag with Prometheus
Alert on MySQL replica lag, stopped replication threads and missing Seconds_Behind_Master values, so stale reads get caught before users report them.
- Fix MySQL ERROR 1040 Too Many Connections
Size mysql max_connections from memory and put ProxySQL in front, so app pods share a bounded pool and peak traffic stops hitting the connection limit.