etcd

Slow Apply Operations Blocking Write Performance

warning
latencyUpdated Jan 25, 2026

High etcd_server_slow_apply_total count indicates that applying committed entries to the state machine is taking too long (>100ms), typically due to disk latency or large database size.

How to detect:

Monitor etcd_server_slow_apply_total counter and etcd_server_apply_slow metric. Check logs for 'apply request took too long' messages. Correlate with database size and disk fsync latency.

Recommended action:

Investigate disk IO performance using iostat. Check database size and run compaction/defragmentation if needed. Review large key-value entries that may slow down applies. Consider moving to faster storage or increasing IOPS allocation.