Integration Incident Postmortem Template
In an order integration incident, restored processing can coexist with missing or duplicated records. A postmortem should account for the records that moved, the ones recovered after detection and the exceptions that still have an owner. The TechSupport Team template keeps those facts alongside the timeline, contributing conditions and prevention work.
Suggested citation · TechSupport Team, “Integration Incident Postmortem Template.” Version 1.0. Reuse under CC BY 4.0.
Copy the postmortem
| Section | What to capture |
|---|---|
| Impact | Affected flow, time range, customer or tenant count where known, and the total number of affected records. |
| Timeline | First detection, impact window, mitigation, verification and customer closure, all in one time zone. |
| Contributing conditions | What let the failure occur and what delayed detection. Mark the hypotheses still under investigation. |
| Record reconciliation | Total scoped, applied at discovery, recovered after, still unresolved or parked, and records remaining unknown. Use mutually exclusive states at the cutoff. |
| Follow-up actions | One measurable change per line with an owner, due date and verification step. |
Blank integration incident postmortem
# Integration Incident Postmortem Template
By TechSupport Team · Version 1.0 · CC BY 4.0
Source: https://techsupport-team.com/resources/integration-incident-postmortem-template/
## Impact
Affected flow, time range, customer or tenant count where known, and the total number of affected records.
## Timeline
First detection, impact window, mitigation, verification and customer closure, all in one time zone.
## Contributing conditions
What let the failure occur and what delayed detection. Mark the hypotheses still under investigation.
## Record reconciliation
Total scoped, applied at discovery, recovered after, still unresolved or parked, and records remaining unknown. Use mutually exclusive states at the cutoff.
## Follow-up actions
One measurable change per line with an owner, due date and verification step.
Use the failed order replay ledger to support the record reconciliation section.
Worked postmortem excerpt
Illustrative incident. At 09:20 UTC, operations detected a gap in a shipment feed. Of 28 scoped records, 20 had already applied, six were confirmed missing and recovered through the approved process, and two remained on a partner validation hold at the cutoff. The team confirmed 26 applied and assigned the remaining two to the partner owner. The report links workflow recovery and subsequent customer follow-up as separate milestones.
Prevention example. Add an alert for a partner-level backlog that crosses the next dispatch deadline. The integration owner verifies it with a staging sample before closing the follow-up. Use incident reviews to improve the system and confirm that the assigned work gets done. Google SRE Workbook on postmortems
Get help with your support queue
Source and reuse
Credit TechSupport Team and link to this page when sharing or adapting version 1.0 under CC BY 4.0.