Fleet System Health

Something is on fire — a signal is breached.

CRITICAL

Refreshed: 2026-07-08 20:12 UTC Fleet data: 0m ago Cadence: 15m · grey > 30m stale Delivery & Work-in-flight →

Delivery velocity OK
≥100/wk
merged PRs (7d) across fleet (≥ — a child hit the 100-PR query cap) · lead-time p50 0.35083333333333333h
updated 0m ago
DORA detail →
CI pipeline health (fleet) CRITICAL
74%
lana 74% · p95 7m · keel: no data (sidecar absent)
no trend updated 0m ago
CI detail →runbook —
Open incidents (on fire?) OK
0
no active incidents (no open `incident` issues across fleet)
no trend updated 0m ago
incident issues →
Change failure rate OK
0.0%
0 revert PRs / 100 merged (7d) · title-heuristic
no trend updated 0m ago
DORA detail →
MTTR NO DATA
—
no resolved incidents (MTTR needs closed `incident` issues)
no trend updated no data
resolved incidents →
CI cost vs baseline OK
0%
0 Actions min this period vs 50000 baseline
no trend updated 0m ago
cost detail →
lana app health NO DATA
no data
Sentry not configured (SENTRY_AUTH_TOKEN absent)
no trend updated no data
Sentry detail →
● OK — in threshold & fresh ◆ WARN — out of threshold / stale ▲ CRITICAL — breach (open incident / threshold breach) □ NO DATA — source absent (never assumed green)

Worst-child rollup — a fleet tile shows its least-healthy product, never the average. Golden-signals / DORA / RED grounded. Colour pairs with a shape for colourblind legibility. Incidents are tracked via the incident label (open = detected, close = restored); MTTR & open-incidents derive only from it, never P0/P1. Thresholds are calibratable: CI success green>95 / amber 85-95 / red<85 · CI cost green<110% / amber 110-130 / red>130 · change-fail green<5 / amber 5-15 / red>15 · MTTR green<24h / amber 24-72h / red>72h · open incident → red.