What are DORA Metrics
DORA metrics are four key indicators developed by the DevOps Research and Assessment (DORA) team, now part of Google Cloud. They measure software delivery performance and operational reliability based on data from thousands of engineering organizations worldwide.
The Four DORA Metrics
Deployment Frequency measures how often code reaches production. Performance tiers: Elite (multiple daily), High (daily to weekly), Medium (weekly to monthly), Low (less than monthly). Lead Time for Changes tracks time from code commit to production execution. Elite performers: under one day; High: one to seven days; Medium: one to four weeks; Low: over a month. Change Failure Rate is the percentage of deployments causing production failures requiring rollback or hotfix. Elite: 0–15%; High/Medium/Low: all 16–30%. Mean Time to Restore Service (MTTR) measures recovery duration from production failure detection to service restoration. Elite: under one hour; High: under one day; Medium: one to seven days; Low: over one week. Elite teams recover 168 times faster than low performers.
Why DORA Metrics Matter
DORA metrics are research-backed, derived from statistical analysis of tens of thousands of professionals across organizations, validated over a decade and published in Accelerate (2018). They are outcome-focused, emphasizing business-relevant metrics rather than activity metrics like lines of code. Research demonstrates elite teams deploy more frequently while maintaining superior reliability: speed and stability reinforce rather than compete with each other. Teams demonstrating stronger delivery performance report improved profitability, market share, and employee satisfaction.
How to Measure DORA Metrics
Deployment Frequency: extract production deployment counts from CI/CD platforms (GitHub Actions, GitLab CI, Jenkins, ArgoCD), excluding staging/development environments. Lead Time: calculate elapsed time between main branch merge and production deployment. Change Failure Rate: track deployments resulting in rollbacks, hotfixes, or incidents, dividing incidents by total deployments. MTTR: pull data from incident management systems, capturing both detection and recovery timestamps. Automation platforms include Sleuth, LinearB, Haystack, and the open-source Four Keys project.
Common Pitfalls and How AI Improves DORA Metrics
Common pitfalls include optimizing a single metric at the expense of others, manipulating metrics through gaming, comparing cross-context teams, and overlooking qualitative factors. AI tools improve DORA metrics directly: AI agents automating root cause analysis compress diagnosis from hours to minutes, improving MTTR. AI pattern analysis identifies high-risk change types, improving change failure rates. Confidence in operational safety nets encourages teams to deploy more frequently rather than batching risky large releases.
What to remember
- 1DORA metrics comprise four research-validated indicators: deployment frequency, lead time for changes, change failure rate, and mean time to restore service
- 2Elite performers deploy on-demand, complete changes in under one day, maintain change failure rates below 15%, and restore service within one hour
- 3Integrated tracking prevents metric gaming; improving one metric through performance degradation elsewhere defeats the framework's purpose
- 4Speed and stability represent mutually reinforcing outcomes rather than competing priorities according to DORA research
- 5AI tools directly enhance MTTR and indirectly support improvements across all four metrics
Frequently asked questions
What are the four DORA metrics?
Deployment Frequency (deployment cadence), Lead Time for Changes (commit-to-production duration), Change Failure Rate (deployment failure percentage), and Mean Time to Restore Service (failure recovery duration).
What does DORA stand for?
DevOps Research and Assessment. Originally independent research led by Dr. Nicole Forsgren, Jez Humble, and Gene Kim, acquired by Google Cloud in 2018. The team publishes annual State of DevOps Reports.
What differentiates elite from low DORA performers?
Elite teams deploy multiple daily with sub-day lead times, 0–15% change failure rates, and sub-hour restoration. Low performers deploy monthly or less, exceed month-long lead times, and require week-plus restoration periods.
How do I begin measuring DORA metrics?
Extract deployment data from CI/CD platforms (GitHub Actions, GitLab, Jenkins). Pull incident data from incident management systems. Automation options include open-source Four Keys or commercial platforms offering DORA dashboards.
Are DORA metrics still relevant in 2026?
Yes. They remain the industry standard for delivery and operational performance measurement. The 2023 DORA report added organizational context nuance while confirming core metric validity.
Can DORA metrics be manipulated?
Yes. Examples include deployment splitting, incident reclassification, or premature ticket closure. Tracking all four metrics simultaneously makes gaming difficult since artificial improvements typically degrade other metrics.
How do DORA metrics relate to MTTR?
Mean Time to Restore Service is one of four DORA metrics and the most operationally relevant. DORA research identifies it as the strongest overall performance predictor.
Are DORA metrics part of Agile frameworks?
Not explicitly part of Agile/Scrum specifications, though commonly adopted by Agile teams. They complement rather than replace traditional Agile metrics like velocity.
What is the DORA report?
Annual State of DevOps Report analyzing survey responses from thousands of engineering professionals, examining relationships between delivery performance and organizational outcomes. Published since 2014.
Are DORA metrics vanity metrics?
No. They drive engineering investment decisions and correlate strongly with business outcomes (profitability, market share, employee satisfaction). Collective tracking prevents gaming without actual improvement.
See it in action. No slides.
NeuBird AI compresses incident investigation from hours to minutes: autonomous root cause analysis, with zero manual triage.