Replication lag is the delay between leader write and follower update:
Causes:
- Network latency between nodes
- Follower under heavy load
- Large transactions
Problems:
- Read-your-writes inconsistency
- Monotonic read violations
- Causality violations
Solutions:
- Read from leader for critical reads
- Track replication position, route reads to caught-up followers
- Use synchronous replication for critical data
- Implement read-after-write guarantees
Monitor replication lag. Alert when it exceeds your consistency SLA.