Technologies/PostgreSQL/postgresql.replication.replay_lag

PostgreSQLMetric

postgresql.replication.replay_lag

Replication replay lag

Dimensions:None

Available on:Native (1)

Datadog (1)

Interface Metrics (2)

Native

replay_lag

replay_lag statistic from pg_stat_replication

Dimensions:None

Datadog

postgresql.replication.wal_replay_lag

Time elapsed between flushing recent WAL locally and receiving notification that this standby server has written, flushed and applied it. This can be used to gauge the delay that synchronous_commit level remote_apply incurred while committing if this server was configured as a synchronous standby. Only available with postgresql 10 and newer.

Dimensions:None

Sources

replay_laggithub.com

postgresql.replication.wal_replay_laggithub.com

Technical Annotations (30)

Configuration Parameters (8)

wal_keep_segmentsrecommended: increase from default

number of WAL segments to retain in pg_xlog directory

max_wal_sendersrecommended: increase from default

maximum concurrent WAL sender processes allowed

archive_moderecommended: on

enable WAL archiving to preserve segments

archive_commandrecommended: configured to alternate path

command to archive WAL files to backup location

max_replication_slotsrecommended: 10

maximum replication slots for lag tracking

wal_compressionrecommended: on

reduces WAL size and network transfer for replication

max_standby_streaming_delayrecommended: 30s

maximum time standby waits before canceling queries during replay

hot_standby_feedbackrecommended: on

prevents query cancellations on standby by informing primary

CLI Commands (7)

SELECT client_addr, state, sent_lsn, write_lsn, flush_lsn, replay_lsn, pg_wal_lsn_diff(sent_lsn, replay_lsn) AS replay_lag_bytes FROM pg_stat_replication;

diagnostic

SELECT client_addr, state, sent_lsn, write_lsn, flush_lsn, replay_lsn, pg_wal_lsn_diff(sent_lsn, replay_lsn) AS lag_bytes FROM pg_stat_replication;

monitoring

SELECT now() - pg_last_xact_replay_timestamp() AS replication_lag;monitoring

ALTER SYSTEM SET max_wal_senders = 10;remediation

ALTER SYSTEM SET wal_keep_segments = 64;remediation

ALTER SYSTEM SET max_standby_streaming_delay = '30s';remediation

ALTER SYSTEM SET hot_standby_feedback = on;remediation

Technical References (15)

WALcomponentWrite-Ahead Loggingconceptpg_xlogfile pathcheckpointconceptpostgres_exportercomponentpgwatchcomponentpg_stat_replicationcomponentWAL applicationconceptfailoverconceptstreaming replicationconceptWAL sendercomponentreplication slotcomponentWAL replayconceptreplication lagconceptclustering architecturesconcept

Related Insights (9)

WAL segment deletion causes replication lag and requires replica rebuildcritical

▸

Replication lag causes replica data staleness and read inconsistencieswarning

▸

Replication lag triggered by CPU saturation from bad query planscritical

▸

Replication lag causes stale replica datacritical

▸

Missing data on UPDATE/DELETE operations silently skipped in logical replicationwarning

▸

Replication lag causing stale reads on standbywarning

▸

Replication lag causes stale reads on replicaswarning

Streaming replication lag between primary and replica nodes increases, causing replicas to serve stale data. High lag indicates replica cannot keep up with primary write volume, network issues, or resource constraints on replica.

▸

Replication lag threatens high availability and data consistencywarning

▸

Long-running queries on read replica with hot_standby_feedback cause periodic CPU spikes on primarywarning

▸