Deployment frequency measures how often your team ships to production. em derives this from GitHub Actions workflow run history — no instrumentation or third-party integrations required.
A deployment is a successful GitHub Actions workflow run (conclusion == "success") for a workflow file that you have configured as a deployment workflow. Failed runs, cancelled runs, and skipped runs are not counted.
You configure which workflows represent deployments per team:
em metrics github configThis lets you map one or more repositories to one or more workflow filenames. For example, a team might count successful runs of deploy-production.yml in their API repo and release.yml in their frontend repo.
Deployments are counted in 7-day buckets anchored backward from the report end date (matching the throughput period logic). This means the most recent bucket ends today, and older buckets represent complete calendar weeks.
Output includes:
- Total successful workflow runs in the date range
- Deployments per week (total ÷ number of weeks)
- Per-week breakdown for the trend chart
Deployment frequency is one of the four DORA metrics. The industry benchmarks for elite performing teams is multiple deployments per day; high performers deploy at least once per week. The weekly view in em makes it easy to spot regressions (a sudden drop in deploy frequency often signals a process problem or a difficult release period).
| Config key | Description |
|---|---|
github.org |
GitHub organization name |
teams.<name>.github.slug |
Team slug (used to scope queries) |
teams.<name>.github.workflows |
Map of repo → [workflow filenames] |
The GitHub API token is stored in the system keychain under the em namespace. It requires at least repo scope to read Actions workflow run history.