DORA Metrics Dashboard

Four-key metrics for software delivery performance · last 56d · data: live · generated 2026-05-10T00:35:07Z ·

Summary

Deploy Freq
250
Elite (on-demand)
merges/week avg · 500 total
Lead Time (p50)
—
—
PR open → merge · n=0
Change Failure Rate
0%
Elite (< 5%)
0 revert PRs / 98 deployed
MTTR
—
No data yet
Will populate post-launch

1. Deployment Frequency

Merges (commits) to main per week — higher is better for Elite performers.

Weekly commit trend (56d)
250 merges/week
Source: gh api repos/{repo}/commits?sha=main&since=… · counts all commits including direct pushes · DORA Elite threshold: on-demand (multiple times/day) · lookback: 56d.

2. Lead Time for Changes

Time from PR opened to PR merged — median and p95 over the lookback window.

Median lead time (p50)
—
No merged PR events in pr-events.jsonl yet. Lead time populates as PRs are journalled by journal-pr-event.sh (#583).
DORA Elite: < 1h · High: < 1d · Medium: < 1w
p95 lead time
—
Sample size: 0 merged PRs
Lookback: 56d
Source: claudedocs/dispatches/pr-events.jsonl · field: merged_at - created_at

3. Change Failure Rate

Percentage of changes that result in a revert or degraded service — lower is better.

Weekly revert count & rate (56d)
0% failure rate
not enough data
Weekly revert PR count · 0 revert(s) detected of 98 forward merges
Detection method: PR title starts with Revert (case-insensitive) · limitation: silent hotfixes not captured · DORA Elite: < 5% · High: 5–10% · Low: ≥ 15%.

4. Mean Time to Restore (MTTR)

Time from incident start to full service recovery — lower is better.

Will populate post-launch. No incidents have been recorded yet — this is expected for a pre-launch system.

When incidents are tracked, populate by logging events with fields: {"event": "incident_opened", "id": "INC-001", "ts": "…"} and {"event": "incident_resolved", "id": "INC-001", "ts": "…"} to claudedocs/incidents.jsonl. The dashboard will compute resolved_ts - opened_ts and report median + p95 MTTR.
DORA Elite: < 1h · High: < 1d · Medium: < 1w · Low: > 1w.

DORA Band Reference

Metric Elite High Medium Low
Deploy Frequency On-demand (> daily) Weekly Monthly < monthly
Lead Time < 1 hour < 1 day < 1 week ≥ 1 week
Change Failure Rate < 5% 5–10% 10–15% ≥ 15%
MTTR < 1 hour < 1 day < 1 week ≥ 1 week
Source: DORA 2023 State of DevOps Report — Accelerate: State of DevOps.

Source: GitHub Actions commit log + claudedocs/dispatches/pr-events.jsonl · rendering script: scripts/generate-dora-dashboard.sh · raw data: data.json · issue #773 · refreshed daily by .github/workflows/dora-dashboard-update.yml.

Known limitations: Deployment Frequency counts all commits to main (not filtered to PR merges only) · Change Failure Rate detects reverts by PR title heuristic only · MTTR requires post-launch incident data · Lead Time requires journal-pr-event.sh (#583) to be running.