Metrics · Updated 2026-09-17
Deployment frequency
How often a team successfully releases to production — a DORA metric, and a proxy for how small and safe its changes are.
What it is
Deployments per day, week or month. Taken alone it is close to meaningless; taken with change failure rate it is one of the better indicators of engineering health, because shipping often and safely requires most of the underlying practices to be working.
Why frequency and batch size are the same conversation
Teams that deploy rarely deploy large batches, and large batches fail in ways that are hard to diagnose — when forty changes ship together, finding which one broke production is archaeology. Frequent small deployments make the culprit obvious and rollback cheap.
That connects to review directly: small pull requests are reviewed better, merged faster and deployed sooner. A review process that encourages large batches suppresses deployment frequency no matter what the pipeline can do.
Why it matters when you are evaluating
It is a poor headline metric for a review tool — nobody deploys more often because a bot commented on their diff. It is a useful guardrail: if frequency drops after you add a blocking check, the check is costing more than it returns.
Common mistakes
- Counting deployments of anything, including config-only no-ops, to inflate the number.
- Chasing frequency without watching stability.
- Comparing across teams with different service architectures.