What goes in a DORA metrics dashboard in Metabase?

A DORA metrics dashboard tracks the four keys of software delivery performance — deployment frequency, lead time for changes, change failure rate, and time to restore service — so throughput and stability sit side by side. Build it from CI/CD, version-control, and incident data synced into a database — see GitHub or GitLab for the connection.

What does a DORA dashboard look like?

Here's the layout this guide builds: the four keys as gauges across the top row, then the trend behind each one — deploys by service, lead-time distribution, change failure rate — and finally the incidents that drive time to restore. Read it monthly, and compare the direction of travel rather than the absolute numbers.

Which cards belong on a DORA dashboard?

The four keys (top row)

• Deployment frequency
• Lead time for changes
• Change failure rate
• Time to restore service (MTTR)

Trends & context

• Deployment frequency by week and by service
• Lead-time-for-changes distribution (median and p90)
• Change failure rate by week
• Incidents and recovery time by service (table)

What data does a DORA dashboard need?

• A deployments table with timestamp, service, and a failure/rollback flag.
• Commit and PR timestamps to compute commit-to-deploy lead time.
• An incidents table with start and resolve times for MTTR.

How do you build a DORA dashboard?

1. Sync deploys and version-control data into a database (GitHub or GitLab), plus incidents from your on-call tool.
2. Model deployments, changes, and incidents onto a shared schema.
3. Compute each of the four keys; show throughput and stability together.
4. Add filters for service, team, and date range.

Example card SQL

-- Deployment frequency and change failure rate by week.
SELECT
  date_trunc('week', d.deployed_at)                       AS week,
  COUNT(*)                                                AS deploys,
  COUNT(*) FILTER (WHERE d.caused_failure)                AS failed_deploys,
  ROUND(100.0 * COUNT(*) FILTER (WHERE d.caused_failure)
        / NULLIF(COUNT(*), 0), 1)                         AS change_failure_rate_pct
FROM deployments d
GROUP BY date_trunc('week', d.deployed_at)
ORDER BY week;

Related

Integrations

Analytics

Dashboards

Metrics

FAQ