Replication lag
The delay between a write committing on the primary database and that write appearing on a replica. An asynchronous replica applies the primary's log after the fact, so the lag grows with write volume, long-running queries on the replica, and network delay. Reads from a lagging replica return old data.
Data
In Quorum#
A Read Replica lags 20 ms plus 2 ms for every write per second, and the share of its reads that come back stale equals the lag in seconds, up to 15%. The Replication lag Incident multiplies the lag by 10, up to 50% stale reads. Stale reads fail API reads and checkouts, while page views accept them; Read-your-writes routing sends the reads that care to the primary.
Read next#
Read-your-writes consistency, Read replica, Synchronous replication, RPO and RTO