INC178379836811310

P3newai_assisted

Nightly backup job failed for orders-db snapshot

Assembled Context — bounded to last 30 minutes

Rolling logs

[backup-worker] ERROR snapshot job orders-db-nightly failed: connection reset

[backup-worker] INFO retry attempt 1/3 scheduled

[backup-worker] WARN last successful backup 26 hours ago (SLA 24h)

Recent commits

(none in lookback window)

Similar past incidents

INC17807673821144 · storage/P3 · 35d ago — Nightly backup job failed for orders-db snapshot

INC177853051382217 · storage/P3 · 60d ago — Nightly backup job failed for orders-db snapshot

INC17837398513061 · storage/P3 · 9999d ago — Nightly backup job failed for orders-db snapshot

Diagnosis Agent

ran in parallel

Backup worker unable to establish a stable connection to the storage backend, causing snapshot creation to fail.

[backup-worker] ERROR snapshot job orders-db-nightly failed: connection reset
[backup-worker] WARN last successful backup 26 hours ago (SLA 24h)
Similar past incidents classified as storage failures (INC17807673821144, INC177853051382217, INC17837398513061)

Assessed: P2 / storage

Self-reported confidence: 90% (audit-only — not used by the gate)

openai/gpt-oss-120b · 2707ms

Severity / Routing Agent

ran in parallel

P3 / storage

Routed to: storage-infrastructure-team

The failure of a nightly backup job beyond the SLA represents a risk to data recoverability, but it is not a production outage. It matches historical incidents classified as P3 for storage. The connection reset suggests a transient infrastructure or network issue during the backup process.

Self-reported confidence: 95% (audit-only — not used by the gate)

gemini-flash-lite-latest · 912ms

Scenario Matcher

deterministic, pre-LLM

Backup job failure

Transient backup job failures are commonly resolved by restarting the backup service before escalating a persistent failure.

Match score: 80% · Recommended action: restart_service

Matched tags: backup, snapshot

Confidence Gate

Computed from 4 measurable signals — not the model's self-reported confidence.

Severity match rate50%

Runbook coverage80%

Data completeness67%

Recency61%

Computed confidence (v1)65%

Auto-execute threshold: 75%