Replication lag causes stale replica data

Replication lag above a few megabytes under normal load means the replica is falling behind the primary. Sustained lag often indicates the replica is under-provisioned or long-running queries are blocking WAL replay. Threshold: alert on lag > few MB.

Check replica server resources (CPU, I/O, memory). Identify long-running queries on the replica blocking WAL replay. Verify network bandwidth between primary and replica. Consider scaling up replica hardware or optimizing query workload on replica.