Know which pipelines are
quietly breaking.
Deviera watches every CI run across all your repos, detects flaky tests and broken pipelines, scores each repo's health, and creates Linear or Jira issues the moment something goes wrong — automatically.
/screenshots/TODO-ci-intelligence.webp
| Repository | Status | Failure rate | Flaky | Runs | Saved |
|---|---|---|---|---|---|
| api-service | broken | 38% | 4 | 120 | 42 min |
| frontend | flaky | 14% | 7 | 210 | 31 min |
| auth-service | stable | 2% | 0 | 85 | 0 min |
| data-pipeline | unstable | 21% | 2 | 64 | 18 min |
Repo Health Scores
0 = critical · 100 = perfect/screenshots/TODO-repo-health.webp
Flaky test detection
Identifies tests that alternate between pass and fail on the same commit — the silent killer of CI confidence.
Failure rate trending
Track failure rates over 7, 14, or 30 days. Spot repos trending toward instability before they break.
Repo health scores
A single 0–100 health score per repo, factoring in failure rate, flakiness, and recent trend direction.
Auto issue creation
When CI breaks, Deviera creates a Linear or Jira issue automatically — with repo, branch, run ID, and failure context.
Cross-repo dashboard
All your repos in one view. Sort by health score, failure rate, or flaky detections to triage the worst offenders first.
Cooldown & dedup
One issue per failure streak, not one per run. Configurable cooldown prevents alert fatigue on long-running outages.
How it works
01
GitHub webhook
Every push to a status check triggers a webhook. Deviera receives the run result in real time.
02
Pattern analysis
Deviera checks failure rate, flaky patterns (alternating pass/fail), and streak length against your configured thresholds.
03
Automated action
A Linear issue is created with full context. Cooldown prevents duplicate issues for the same ongoing failure.
< 30s
Time from CI fail to Linear issue
60 min
Default cooldown between issues
7-day
Flaky detection window
Connect GitHub, enable a CI automation, and let Deviera watch your pipelines around the clock.