Incidents

Which playbook should I open?

Payment incidents are not interchangeable. Classify by operational signal and incident class, then open the playbook and reference that match your plane of failure. When uncertain, start with payment incident triage.

Incident taxonomy

Six classes cover the majority of production payment operations incidents. Each class maps to a primary playbook—secondary playbooks handle adjacent failure modes.

Signal → playbook routing

Direct routing when the signal and class are already clear. The summary table uses playbook names; linked routes follow in the list below.

Start with payment incident triage when class is unclear; use rows below for direct routing.

SignalIncident classOpen playbookSupporting reference
Webhook recency lagWebhook / ProviderPayment incident triageWebhook delivery model
Webhook recency lag (sustained, API errors)ProviderProvider outage responseProvider retry semantics
Checkpoint lag (detection → Paid)Detection / SettlementDelayed settlement recoverySettlement checkpoint model
Checkpoint lag (policy confirmation)SettlementConfirmation escalationConfirmation policy matrix
Exception queue depth risingReconciliationException queue triageReconciliation state model
Reconciliation drift (repeat matcher failure)ReconciliationReconciliation closeReconciliation state model
Provider latency / API errorsProviderProvider outage responseProvider retry semantics
Payout review backlogPayoutMerchant payout reviewLedger transitions
Duplicate payment / replay side effectsWebhook / DetectionDuplicate investigationWebhook delivery model
Amount variance (under/over)ReconciliationUnder/overpayment handlingReconciliation state model

Open payment incident triage to classify signals, assign incident class, and route to the correct playbook without skipping evidence collection.