DORA's four keys are deployment frequency, change lead time, change failure rate and failed deployment recovery time, and deployment rework rate is the fifth. In DORA's 2025 report, more AI adoption went with higher delivery throughput and still more instability.
CI wall time 12.4mRuns per PR 2.5Test / verify activity 18h / week
Approve → merge28m per PR
The largest delay comes after approval: the work is ready, but the merge is still waiting.
Give the agent a clear merge rule: merge when approval and all required checks are green.About this sample
Sample from Zest’s pipeline report · Stage medians overlap and do not add up. Test / verify activity is a weekly session-flow estimate, not pipeline wall time.
↓ Scroll to explore
USED BY ENGINEERING TEAMS AT
What are DORA metrics?
DORA metrics are the software delivery measures from DORA (DevOps Research and Assessment), a research program run by Google Cloud. The original four keys are:
Deployment frequency: how often you deploy to production, or the time between deployments.
Lead time for changes: how long a change takes from commit to running in production. DORA now calls it change lead time.
Change failure rate: the share of deployments that need immediate intervention, such as a rollback or a hotfix.
Time to restore service (MTTR): how long it takes to recover from a failure. DORA has replaced it with failed deployment recovery time, the time to recover from a deployment that fails and needs immediate intervention.
Throughput, instability, and the fifth metric
DORA groups its metrics in two. Throughput (change lead time, deployment frequency and failed deployment recovery time) measures how many changes move through the system. Instability (change fail rate and deployment rework rate) measures how well deployments go.
Deployment rework rate is the newest of the five: the share of deployments that are unplanned and happen because of an incident in production.
How AI coding changes DORA metrics
DORA's 2025 report, State of AI-assisted Software Development, surveyed nearly 5,000 technology professionals, and 90% of them use AI at work. Unlike in 2024, more AI adoption now goes with higher software delivery throughput, but still with more delivery instability.
DORA's reading is that AI amplifies what a team already has. Without good tests, mature version control and fast feedback, more changes mean more instability. Coding agents make each change cheaper to write, so the change volume DORA warns about is exactly what they raise.
What to measure next to DORA on an AI-assisted team
DORA metrics start at the commit, and DORA is only one of the developer productivity metrics frameworks. On an AI-assisted team, much of the work happens before the commit, in the coding-agent sessions that produce the change. That is the part Zest measures:
Which coding agents, models and skills each engineer uses: Claude Code, Cursor, Codex, GitHub Copilot Chat and more.
Pull requests opened and merged, PR throughput and PR cycle time, from the Zest GitHub App.
Which sessions led to which pull requests.
What Zest does not measure
Zest does not calculate DORA metrics. It doesn't ingest deployments or incidents, so deployment frequency, change failure rate and recovery time stay with your CI/CD and incident tools. Use DORA for delivery, and Zest for the AI coding work that feeds it.
TRY IT ON YOUR TEAM
See what your team ships with AI.
Starter is free (3 seats, no credit card). Pro is $49 a month with 5 seats and Max is $149 a month with 10 seats — priced per workspace, not per engineer, with extra seats at $19 a month.
Deployment frequency, lead time for changes (now change lead time), change failure rate, and time to restore service, which DORA has replaced with failed deployment recovery time.
Is MTTR still a DORA metric?+
DORA replaced mean time to restore with failed deployment recovery time: the time it takes to recover from a deployment that fails and requires immediate intervention.
What is the fifth DORA metric?+
Deployment rework rate: the ratio of deployments that are unplanned and happen as a result of an incident in production. DORA groups it with change fail rate as a measure of instability.
Does AI coding improve DORA metrics?+
Throughput, yes; stability, not yet. In DORA's 2025 research, more AI adoption went with higher delivery throughput and still with more delivery instability. DORA's advice: good tests, mature version control and fast feedback.
Winding / Agentic loop
