DevOps Research and Assessments (DORA) metrics are used to understand software engineering team productivity. It focuses on measuring metrics on the Outcome and Impact levels of the Productivity Framework.
This framework aims to measure and categorize team performance into four levels:
- Elite
- High
- Medium
- Low
There are four metrics in this framework that attempt to measure two core areas of DevOps:
| Area | Metric |
|---|---|
| Speed | Deployment frequency |
| Mean Lead Time for Changes | |
| Stability | Change Failure Rate |
| Mean Time to Resolution (MTTR) |
Deployment Frequency
Indicates how frequency can a team successfully deploy code to production or releases to end users.
This feels poorly defined. Is this a measure of throughput in the sense that if we have high throughput we have higher deployment frequency? Or purely about how fast can CI/CD deploy out changes — the faster it is, the more frequent the deployment.
| Performance | Benchmark |
|---|---|
| Elite | Multiple times a day |
| High | One per week to one per month |
| Medium | One per month to one every six months |
| Low | Less than one every six months |
Mean Lead Time for Changes
Definition
How long does it take on average to complete a feature from start to finish? From customer feedback to deployment.
| Performance | Benchmark |
|---|---|
| Elite | Within one hour |
| High | One day to one week |
| Medium | One month to six months |
| Low | More than six months |
Change Failure Rate
Definition
Percentage of deployments causing a failure in production that require an immediate fix.
Examples of such failures include service degradation or an outage.
| Performance | Benchmark |
|---|---|
| Elite | 0%-15% |
| High | 16%-30% |
| Medium | 30%-50% |
| Low | 50%+ |
Mean Time to Recover
Definition
Average time taken to recover from a customer-facing incident.
| Performance | Benchmark |
|---|---|
| Elite | Within one hour |
| High | Within one day |
| Medium | One day to one week |
| Low | More than six months |