Cross-Region Stale Read Window
A multi-region deployment routes reads to the nearest regional replica. Under write surges or inter-region network degradation, the regional replica accumulates replication lag. Reads from the lagging region return state that contradicts recent writes: violating read-after-write consistency. If lag reaches the WAL retention boundary, the replica cannot resync incrementally and requires a full base backup across regions.
Primary region (us-east) + replica region (eu-west) + global read router + regional CDN
Degradation Replay
Nominal: Replicas Synchronized
All regional replicas within acceptable lag threshold
- ·Cross-region replication lag < 500ms: imperceptible to users
- ·Regional reads consistent with primary within RTT + apply time
- ·WAL streaming healthy: no segment accumulation on primary
- ·All regional replicas receiving and applying WAL continuously
- ·Read router directing regional traffic to local replica
- ·Inter-region bandwidth utilization normal
- !Users in all regions see consistent data within acceptable eventual-consistency window
Operational simulation model only, not a production forecast. Degradation stages are derived from structured operational knowledge, not measured telemetry. Do not use for capacity planning or incident response.
Run With Your Parameters
Adjust the parameters below to see how metric values shift across degradation stages. Formulas are deterministic: same inputs always produce the same output.
Simulation Parameters
Network RTT between primary and replica region
Computed Degradation Stages
All regional replicas within acceptable lag threshold
Write rate on primary increases or inter-region network latency rises
Lag exceeds 5-second SLA; read router begins rerouting to primary
Lag so large that primary WAL segments needed by replica are at risk of rotation
Write burst subsides; replica catching up; lag decreasing
Interpretation
Parameterized simulation for Cross-Region Stale Read Window completed.
See computed stage metrics for details.
Refer to stabilization patterns and recovery patterns for mitigation guidance.
Parameterized simulation: not a production forecast. Values derived from deterministic formulas applied to your parameters. Do not use for capacity planning or operational decisions without validation.
Propagation Model
WAL generation rate exceeds inter-region transfer rate; replica accumulates lag at (write_rate - transfer_rate) bytes/sec
Stabilizes: Lag self-heals once write burst subsides and replica apply rate exceeds new write rate
Read router continues routing to lagging replica; users in affected region see stale data disproportionate to lag depth
Stabilizes: Stale reads stop once replica LSN overtakes user write LSN: requires lag to fully clear
Read router redirects regional reads back to primary; primary absorbs both write workload and inter-region read traffic: OLTP latency rises
Stabilizes: Primary load normalizes once replica lag clears and read traffic returns to replica
Recovery Patterns
Self-healing after write burst completion
5-30 minutes after burst subsides, depending on lag depth- ·During recovery, primary handles all reads: OLTP latency elevated
- ·Users in affected region experience elevated latency until replica catches up
- !If WAL segments rotate before replica catches up, full resync required (hours)
Full cross-region pg_basebackup resync
2-8 hours depending on database size and inter-region bandwidth- ·Consumes inter-region bandwidth during resync: further slows any ongoing WAL streaming
- ·Replica unavailable during resync: all reads hit primary
- !Resync window must complete before next write burst: incomplete resync restarts from scratch
Operational Summary
Cross-region replication lag accumulates when primary WAL generation exceeds inter-region transfer capacity. The stale read window grows proportionally to lag depth, causing read-after-write violations for users in the lagging region. The critical risk is WAL segment rotation: if the primary recycles segments the replica hasn't applied, the replica must resync via pg_basebackup: a multi-hour operation. Prevention requires sufficient wal_keep_size, inter-region bandwidth headroom, and lag-aware read routing.