RedisMetricredis.replication.repl_backlog_size
Replication backlog size
Summary
Size in bytes of the replication backlog buffer on the master. This buffer stores recent write operations to support partial resynchronization when replicas disconnect temporarily. If a replica's lag exceeds the backlog size, it requires full resynchronization (expensive). As mentioned in the insight about replication backlog sizing, undersized backlogs combined with network instability force frequent full resyncs, impacting performance.
Interface Metrics (3)
Related Insights (2)
Redis replica falling behind master (redis.replication.master_last_io_seconds_ago increasing) continues serving stale cached data with normal hit rates, masking synchronization issues until applications encounter data inconsistencies or master failover fails.
When redis.replication.repl_backlog_size is too small relative to write rate and network latency, replicas require full resync after brief disconnections, causing replication storms and increased master load.