Alert on PostgreSQL WAL archiving backlog before space runs out

Connect archive-failure alerts with WAL capacity and an owned repair procedure.

Source facts

PostgreSQL 17 continues accumulating WAL segments in pg_wal when its archive command or archive library cannot make progress. An empty archive_command can deliberately pause shell-based archiving, but the WAL still accumulates. Repeated archive failures or an archive destination that needs operator intervention therefore become a local capacity problem as well as a recovery problem. If the filesystem holding pg_wal fills, PostgreSQL performs a panic shutdown and remains unavailable until space is made available.

Source: PostgreSQL 17 — Continuous Archiving and PITR.

Applicability

This applies to PostgreSQL 17 continuous-archiving deployments. The behavior described is an archive-backlog condition; it does not identify every possible cause of WAL growth.

DSE recommendation

DSE recommends monitoring archive failure and free space together, with a named responder and an escalation path for an inaccessible or full archive destination. Write a repair procedure that preserves the recovery chain and restores the archiver’s normal progress. Do not make deleting unexplained WAL files the emergency playbook. Include a communication step when recovery coverage is degraded, and keep enough diagnostic information to distinguish a destination problem from a command or permission failure.

Verification

In a disposable environment, introduce a controlled archive-destination failure and observe both the archiver’s error and retained WAL growth. Repair the destination and verify that archiving catches up. Retain the alert timeline and recovery-chain checks; do not reproduce a full production filesystem as the test.

Official reference

PostgreSQL 17 — Continuous Archiving and PITR. Reviewed October 15, 2026.