Webhook backlog runbook

When to activate

Activate this runbook when webhook queue size exceeds threshold and processing latency degrades.

Strategy

• Preserve idempotency before increasing throughput.
• Prioritize critical billing and payment-state events.
• Avoid blind replay without filtering windows.

Procedure

1. 11) Measure backlog

   Quantify delay by event type and organization.

2. 22) Increase capacity

   Scale workers in controlled batches and monitor error rate.

3. 33) Apply controlled replay

   Reprocess with clear tagging and active deduplication.

4. 44) Validate reconciliation

   Confirm convergence between processed events and dashboard final state.

Expected outcome

• Backlog returns to nominal window.
• No duplicated business side effects.
• Complete audit trail for reprocessed events.