Reliability Analytics Schemas — Operating Quantum Computers

This appendix provides starter schemas for fleet-health analytics, reliability reviews, and health-driven admission control.

Diagram loads as you read

Reliability Analytics Schemas · Figure 1

View diagram source

flowchart TB
    Schema[Reliability schema set] --> Snapshot[Health snapshot]
    Schema --> Component[Health component]
    Schema --> Incident[Incident]
    Schema --> Action[Mitigation action]
    Schema --> Review[Reliability review]

Health snapshot

Illustrative listing · yaml

health_snapshot:
  snapshot_id: hs_2026_04_19_001
  target_id: provider.qpu.device
  timestamp: 2026-04-19T12:00:00Z
  target_class: superconducting_gate_model
  health_state: yellow
  aggregate_score: 78
  calibration_age_hours: 31.4
  production_probe_success_rate: 0.91
  missing_telemetry_pct: 0.4
  decision:
    production_admission: restricted
    educational_admission: allowed

Diagram loads as you read

Health snapshot · Figure 2

View diagram source

erDiagram
    HEALTH_SNAPSHOT ||--o{ HEALTH_COMPONENT : contains
    HEALTH_SNAPSHOT ||--o{ ADMISSION_DECISION : drives
    HEALTH_SNAPSHOT }o--|| TARGET : describes

Health component

Illustrative listing · yaml

health_component:
  snapshot_id: hs_2026_04_19_001
  component_name: two_qubit_error_trend
  value: 0.014
  baseline: 0.010
  normalized_score: 62
  state: warning
  evidence_uri: evidence://health/two_qubit_error/run_441

Incident schema

Illustrative listing · yaml

incident:
  incident_id: qi_2026_04_19_002
  category: soft_degradation
  detected_at: 2026-04-19T13:04:00Z
  user_visible: true
  affected_targets:
    - provider.qpu.device
  affected_workload_classes:
    - production_benchmark
  detection_source: fleet_health_policy
  status: investigating
  severity: sev2

Diagram loads as you read

Incident schema · Figure 3

View diagram source

flowchart LR
    Incident[Incident] --> Detection[Detection]
    Incident --> Impact[Impact]
    Incident --> Action[Action]
    Incident --> Review[Review]

Mitigation action

Illustrative listing · yaml

mitigation_action:
  action_id: ma_2026_04_19_004
  incident_id: qi_2026_04_19_002
  action_type: restrict_admission
  target_id: provider.qpu.device
  risk_class: reversible_but_impactful
  approved_by: platform_oncall
  rollback_condition: production_probe_success_rate_gt_0_96_for_3_runs

Weekly reliability review template

Illustrative listing · markdown

# Quantum Reliability Review

Date:
Owner:
Targets reviewed:

## Fleet summary
- Green targets:
- Yellow targets:
- Red targets:

## Material degradations
| Target | State | Workloads affected | Action |
|---|---|---|---|

## Late detections
| Incident | Detection gap | Corrective action |
|---|---|---|

## Policy changes
| Policy | Change | Evidence |
|---|---|---|

## Backlog
| Item | Owner | Due |
|---|---|---|

Diagram loads as you read

Weekly reliability review template · Figure 4

View diagram source

flowchart TB
    Review[Weekly review] --> Degradation[Degradation list]
    Review --> Late[Late detections]
    Review --> Policy[Policy changes]
    Review --> Backlog[Backlog]

Admission policy fragment

Illustrative listing · yaml

admission_policy:
  workload_class: production_benchmark
  requirements:
    health_state: green
    calibration_age_hours_max: 24
    evidence_required: true
    benchmark_probe_success_rate_min: 0.95
  fallback:
    allow_simulator: true
    allow_qpu: false

Reliability warehouse checklist

Diagram loads as you read

Reliability warehouse checklist · Figure 5

View diagram source

flowchart LR
    Ingest[Ingest] --> Validate[Validate]
    Validate --> Store[Store]
    Store --> Analyze[Analyze]
    Analyze --> Act[Act]
    Act --> Review[Review]

Every health snapshot has a target identifier.
Every aggregate score decomposes into components.
Every user-visible degradation links to impacted workloads.
Every automated action has a policy version.
Every policy change has review evidence.