Radiology Service — Failure Modes
Status: populated Owner: TBD Last updated: 2026-04-18 Companion: Service Template · 03 platform-services · 02 DDD
1. Failure Catalog
| # | Failure | User impact | Detection | Mitigation |
|---|---|---|---|---|
| F-01 | PACS DICOMweb endpoint unreachable | Study metadata not refreshed; viewer launch blocked | rad_pacs_qido_success_rate drops; alert P2 | Show cached metadata; clear error message; do not block chart |
| F-02 | PostgreSQL unavailable | All reads/writes fail | Health probe fails | Failover replica; read-only mode for study list from cache |
| F-03 | NATS unavailable | Events not published; chart not updated on sign | Outbox lag alert | Outbox accumulates; events published on reconnect |
| F-04 | FHIR gateway unreachable on report sign | DiagnosticReport not filed; chart delayed | Outbox lag alert | Retry via outbox; sign action succeeds locally; chart delayed |
| F-05 | Critical finding event not delivered | Urgent finding not notified | No downstream ack within 15 min | Republish via outbox retry; escalation via communication-service |
| F-06 | Viewer token secret rotation | All outstanding tokens invalid | 401 spike on viewer load | Rotation window: dual-key period; brief viewer session interruption |
| F-07 | PACS UID mismatch (wrong patient) | Wrong images shown | Audit log detection; QIDO-RS patient ID check | Require admin mapping override with audit; block auto-display |
| F-08 | Multiple PACS endpoints per facility | Incorrect endpoint selected | Study not found in expected PACS | Facility configuration review; priority ordering for endpoint resolution |