How much replication lag is normal?

Milliseconds to low tens of milliseconds for replicas in the same data center under normal load. It spikes during bulk writes, schema migrations, and replica hardware trouble, sometimes to minutes. Treat lag as a monitored metric with an alert, because every read-from-replica strategy quietly assumes it stays small.

Does replication mean I don't need backups?

No. Replicas replay everything, including your mistakes, so a bad DELETE or dropped table propagates to every copy within seconds. Backups plus point-in-time recovery protect against human error and logical corruption; replication protects against hardware failure. GitLab's 2017 incident is the standard cautionary tale for confusing the two.

When is multi-leader replication actually worth it?

When you genuinely need low-latency writes in multiple regions, or clients that write while offline and sync later, and you are prepared to own conflict resolution. If write latency to one region is tolerable, a single leader with regional read replicas gives you most of the benefit with none of the conflict machinery.

Scalingchapter 3 of 4 · 4 lessons

Database Replication

Copy data across multiple DB nodes for reads and high availability.

← All chapters View in cheatsheet

Your database is one machine. At 03:40 on a Sunday its disk controller fails.

Everything stops, and it stays stopped until somebody wakes up, finds last night's backup, and restores it onto new hardware. That is four hours if the runbook is good, and you have lost every order placed since the backup ran.

Replication is keeping a second machine that already has the data, applying every change as it happens. It also scales your reads and puts data nearer your users, and those are the bonuses. The reason you do it is the Sunday.

The catch is that copies drift, and almost all the engineering here is about how much drift you can stand.

Lessons

4 in this chapter

Leader-Follower BasicsAll writes go through one door. Everything else is copies replaying the leader's log.2 min
Sync vs Async ReplicationDoes the leader wait for a follower before saying done? That decides what a crash can eat.2 min
Replication Lag in PracticeReplicas run behind by design, and users notice the moment they read their own missing write.3 min
Failover and Multi-LeaderPromoting a follower is easy. Fencing the old leader and merging conflicts is the hard part.3 min