Reusable GitHub Composite Action to calculate DORA metrics and upload CSV artifacts.

DoraPulse calculates the four key metrics defined by the DORA (DevOps Research and Assessment) team.

• How often an organization successfully releases to production
• Elite performers: Multiple deployments per day
• Metric shows: Release cadence and batch size

• Time taken from code commit to code running in production
• Elite performers: Less than one hour
• Metric shows: Process efficiency and deployment pipeline health

• How long it takes to restore service after an incident
• Elite performers: Less than one hour
• Metric shows: Incident response effectiveness

• Percentage of changes that lead to degraded service
• Elite performers: 0-15%
• Metric shows: Quality of testing and review processes

• Official DORA Research Program
• Google Cloud's DORA Metrics
• DORA's State of DevOps Reports
• DORA Metrics in GitHub
• Accelerate: The Science of Lean Software - The book that introduced DORA metrics

• dora_metrics.csv: per-repository metrics
• dora_metrics_summary.csv: aggregate summary

• token (required): GitHub token with read access to target repositories
• owner (optional): repository owner for single repository mode, or optional owner/org filter in topic mode
• repo (optional): repository name for single repository mode
• topic (optional): GitHub topic for discovery mode
• days (optional, default 30)
• artifact_name (optional, default dora-metrics-csv)

Provide either:

• owner + repo (single repository mode), or
• topic (topic mode, optional owner filter)

Metrics are calculated from closed pull requests with merged_at inside the selected rolling window (days).

• Deployment Frequency: merged_prs_in_window / days
• Lead Time for Changes (hours): Average of (merged_at - created_at) for merged PRs in window
• MTTR (hours): Average of (merged_at - created_at) only for merged PRs whose title/body contains recovery keywords: fix, hotfix, revert, rollback
• Change Failure Rate (%): (failed_prs / merged_prs_in_window) * 100 A PR is treated as failed if title/body contains any failure keyword: fix, hotfix, revert, rollback, emergency

CSV files:

• dora_metrics.csv: One row per repository with columns: repository, period_days, merged_prs, failed_prs, deployment_frequency_per_day, lead_time_hours, mttr_hours, change_failure_rate_pct
• dora_metrics_summary.csv: Columns: repositories_analyzed, period_days, total_merged_prs, total_failed_prs, avg_deployment_frequency_per_day, avg_lead_time_hours, avg_mttr_hours Weighted failure rate formula: weighted_change_failure_rate_pct = (total_failed_prs / total_merged_prs) * 100

Notes:

• Unmerged PRs are ignored for all four metrics.
• If no merged PRs are found in window, rate/time metrics are emitted as 0.0.

name: DORA Metrics

on:
  workflow_dispatch:

jobs:
  dora:
    runs-on: ubuntu-latest
    steps:
      - name: Run DoraPulse
        uses: chandanrattan/DoraPulse@v1
        with:
          token: ${{ secrets.TOKEN }}
          owner: my-org
          repo: my-service
          days: 30
          artifact_name: my-service-dora

name: DORA Topic Metrics

on:
  workflow_dispatch:

jobs:
  dora-topic:
    runs-on: ubuntu-latest
    steps:
      - name: Run DoraPulse
        uses: chandanrattan/DoraPulse@v1
        with:
          token: ${{ secrets.TOKEN }}
          topic: microservice
          owner: my-org
          days: 30

The action uses scripts/dora_metrics_csv.py.

1. Install dependencies:

pip install .

2. Copy env template and set values:

cp .env.example .env

3. Edit .env and choose one mode:

• Single repository mode: set TOKEN, DORA_OWNER, DORA_REPO and keep DORA_TOPIC empty
• Topic mode: set TOKEN, DORA_TOPIC and optionally DORA_OWNER as filter; keep DORA_REPO empty

4. Run:

python scripts/dora_metrics_csv.py

Optional examples:

# Estimate only (no CSV generation)
python scripts/dora_metrics_csv.py --estimate-only

# CLI override example
python scripts/dora_metrics_csv.py --owner my-org --repo my-service --token <token>

Install with dev tooling:

pip install ".[dev]"

For local execution:

python scripts/dora_metrics_csv.py

Use .env / .env.example variables:

• TOKEN
• DORA_OWNER + DORA_REPO, or DORA_TOPIC
• DORA_DAYS
• DORA_OUT_DIR