Use this appendix for runtime, hardware, data, scientific-claim, and governance incidents.
Standard template
# Incident postmortem: <title>
## Summary
- Incident ID:
- Detected at:
- Resolved at:
- Severity:
- Affected systems:
- Affected claims or datasets:
## Impact
- Users affected:
- Jobs affected:
- Benchmark entries affected:
- Claims affected:
- Decisions affected:
## Timeline
- T0:
- T1:
- T2:
## Root causes
- Technical contributing factors:
- Process contributing factors:
- Detection gaps:
## Corrective actions
- Immediate fix:
- Preventive control:
- Test or monitor addition:
- Owner:
- Due date:View diagram source
flowchart LR
Detect[Detect] --> Stabilize[Stabilize]
Stabilize --> Scope[Scope impact]
Scope --> Analyze[Analyze causes]
Analyze --> Actions[Corrective actions]
Actions --> Verify[Verify]Quantum-specific impact table
| Area | Required question |
|---|---|
| Jobs | Which job IDs, workflow IDs, and sessions were affected? |
| Hardware | Which backend, calibration window, qubits, couplers, or zones were affected? |
| Compilation | Which compiler version, target snapshot, and optimization profile were used? |
| Data | Which raw, processed, and derived datasets are suspect? |
| Claims | Which decks, dashboards, papers, or customer reports used the result? |
| Governance | Did access, quota, export-control, or review policy contribute? |
View diagram source
flowchart TB
Impact[Impact analysis] --> Jobs[Jobs]
Impact --> Hardware[Hardware context]
Impact --> Compile[Compiler context]
Impact --> Data[Data lineage]
Impact --> Claims[Claim usage]
Impact --> Governance[Governance]Claim rollback record
claim_rollback:
claim_id: ""
original_status: active
new_status: provisional | superseded | withdrawn
reason: ""
affected_artifacts:
- uri: ""
type: result_card | dashboard | paper | deck | customer_report
replacement_result: ""
approver: ""
date: ""View diagram source
stateDiagram-v2
Active --> Provisional
Provisional --> Reinstated
Provisional --> Superseded
Provisional --> WithdrawnQuarterly review
View diagram source
flowchart TB
Incidents[Quarter incidents] --> Cluster[Cluster by failure mode]
Cluster --> Trends[Trend analysis]
Trends --> Gaps[Control gaps]
Gaps --> Roadmap[Reliability roadmap]
Roadmap --> Owners[Owners and due dates]Quarterly questions:
- Which incident class repeated?
- Which corrective actions remained open?
- Which actions were too weak?
- Which monitors detected incidents late?
- Which claims required rollback?
- Which architectural changes would remove a failure class?