Skip to main content

Recovery readiness

The Recovery Readiness tab gives an operational summary of crash recovery and data safety posture.

What it covers​

  • Archive health — whether WAL archiving is enabled and functioning
  • Replica usefulness — whether streaming replicas are present, synchronous, and current
  • WAL retention hazards — slot-retained and wal_keep_size-retained WAL
  • Restore prerequisites — checklist of conditions for a viable PITR restore

Crash recovery (RTO)​

pocketPG estimates crash recovery time from max_wal_size and a conservative WAL replay speed assumption (~50 MB/s). The estimate is deliberately conservative — actual speed depends on hardware, WAL content, and checkpoint frequency.

The estimate only covers the WAL replay phase, not the full instance restart time.

Data safety (RPO)​

Data safety is assessed from:

  • Replica sync state — synchronous vs asynchronous
  • Replica write lag — worst-case lag across sync/quorum replicas
  • Slot retention — WAL retained by replication slots
  • Archive status — whether WAL is being archived for PITR

These signals are combined into a directional summary, not a precise RPO number.

Caveats​

pocketPG does not verify:

  • backup integrity
  • restore success
  • backup chain completeness
  • offline WAL file accessibility

These require a tested disaster recovery procedure. The Recovery Readiness tab is an operational summary to help identify gaps, not a validation of DR posture.