跳到主要内容
版本:v0.14

DORA

This document describes everything you need to know about DORA, and implementing this powerful and practical framework in DevLake.

What are DORA metrics?

Created six years ago by a team of researchers, DORA stands for "DevOps Research & Assessment" and is the answer to years of research, having examined thousands of teams, seeking a reliable and actionable approach to understanding the performance of software development teams.

DORA has since become a standardized framework focused on the stability and velocity of development processes, one that avoids the more controversial aspects of productivity and individual performance measures.

There are two key clusters of data inside DORA: Velocity and Stability. The DORA framework is focused on keeping them in context with each other, as a whole, rather than as independent variables, making the data more challenging to misinterpret or abuse.

Within velocity are two core metrics:

  • Deployment Frequency: Number of successful deployments to production, how rapidly is your team releasing to users?
  • Lead Time for Changes: How long does it take from commit to the code running in production? This is important, as it reflects how quickly your team can respond to user requirements.

Stability is composed of two core metrics:

To make DORA even more actionable, there are well-established benchmarks to determine if you are performing at "Elite", "High", "Medium", or "Low" levels. Inside DevLake, you will find the benchmarking table available to assess and compare your own projects.

Why is DORA important?

DORA metrics help teams and projects measure and improve software development practices to consistently deliver reliable products, and thus happy users!

How to implement DORA metrics with Apache DevLake?

You can set up DORA metrics in DevLake in a few steps:

  • Install: Getting Started
  • Collect: Collect data via blueprint
    • In the blueprint, select the data you wish to collect, and make sure you have selected the data required for DORA metrics
    • Configure DORA-related transformation rules to define deployments and incidents
    • Select a sync frequency for your data, save and run the blueprint.
  • Report: DevLake provides a built-in DORA dashboard. See an example screenshot below or check out our live demo. DORA Dashboard

DevLake now supports Jenkins, GitHub Action and GitLabCI as data sources for deployments data; Jira, GitHub issues, and TAPD as the sources for incidents data; Github PRs, GitLab MRs as the sources for changes data.

If your CI/CD tools are not listed on the Supported Data Sources page, have no fear! DevLake provides incoming webhooks to push your deployments data to DevLake. The webhook configuration doc can be found here.

A real-world example

Let's walk through the DORA implementation process for a team with the following toolchain

  • Code Hosting: GitHub
  • CI/CD: GitHub Actions + CircleCI
  • Issue Tracking: Jira

Calculating DORA metrics requires three key entities: changes, deployments, and incidents. Their exact definitions of course depend on a team's DevOps practice and varies team by team. For the team in this example, let's assume the following definition:

  • Changes: All pull requests in GitHub.
  • Deployments: GitHub action jobs that have "deploy" in their names and CircleCI's deployment jobs.
  • Incidents: Jira issues whose types are Crash or Incident

In the next section, we'll demonstrate how to configure DevLake to implement DORA metrics for the aforementioned example team.

Collect GitHub & Jira data via blueprint

  1. Visit the config-ui at http://localhost:4000

  2. Create a blueprint, let's name it "Blueprint for DORA", add a Jira and a GitHub connection. Click Next Step

  3. Select Jira boards and GitHub repos to collect, click Next Step

  4. Click Add Transformation to configure for DORA metrics

  5. To make it simple, fields with a label are DORA-related configurations for every data source. Via these fields, you can define what are "incidents" and "deployments" for each data source. After all data connections have been configured, click Next Step

    • This team uses Jira issue types Crash and Incident as "incident", so choose the two types in field "incident". Jira issues in these two types will be transformed to "incidents" in DevLake.
    • This team uses the GitHub action jobs named deploy and build-and-deploy to deploy, so type in (?i)deploy to match these jobs. These jobs will be transformed to "deployments" in DevLake.

    Note: The following example shows where to find GitHub action jobs. It's easy to mix them up with GitHub workflows.

  1. Choose sync frequency, click 'Save and Run Now' to start data collection. The time to completion varies by data source and depends on the volume of data.

For more details, please refer to our blueprint manuals.

Collect CircleCI data via webhook

Using CircleCI as an example, we demonstrate how to actively push data to DevLake using the Webhook approach, in cases where DevLake doesn't have a plugin specific to that tool to pull data from your data source.

  1. Visit "Data Connections" page in config-ui and select "Issue/Deployment Incoming Webhook".

  2. Click "Add Incoming Webhook", give it a name, and click "Generate POST URL". DevLake will generate URLs that you can send JSON payloads to push deployments and incidents to Devlake. Copy the Deployment curl command.

  3. Now head to your CircleCI's pipelines page in a new tab. Find your deployment pipeline and click Configuration File

  4. Paste the curl command copied in step 8 to the config.yml, change the key-values in the payload. See full payload schema here.

    version: 2.1

    jobs:
    build:
    docker:
    - image: cimg/base:stable
    steps:
    - checkout
    - run:
    name: "build"
    command: |
    echo Hello, World!

    deploy:
    docker:
    - image: cimg/base:stable
    steps:
    - checkout
    - run:
    name: "deploy"
    command: |
    # The time a deploy started
    start_time=`date '+%Y-%m-%dT%H:%M:%S%z'`

    # Some deployment tasks here ...
    echo Hello, World!

    # Send the request to DevLake after deploy
    # The values start with a '$CIRCLE_' are CircleCI's built-in variables
    curl https://sample-url.com/api/plugins/webhook/1/deployments -X 'POST' -d "{
    \"commit_sha\":\"$CIRCLE_SHA1\",
    \"repo_url\":\"$CIRCLE_REPOSITORY_URL\",
    \"start_time\":\"$start_time\"
    }"

    workflows:
    build_and_deploy_workflow:
    jobs:
    - build
    - deploy

    If you have set a username/password for Config UI, you need to add them to the curl to register a deployment:

    curl https://sample-url.com/api/plugins/webhook/1/deployments -X 'POST' -u 'username:password' -d '{
    \"commit_sha\":\"$CIRCLE_SHA1\",
    \"repo_url\":\"$CIRCLE_REPOSITORY_URL\",
    \"start_time\":\"$start_time\"
    }'
  5. Run the modified CircleCI pipeline. Check to verify that the request has been successfully sent.

  6. You will find the corresponding deployments in table.cicd_tasks in DevLake's database.

View and customize DevLake's DORA dashboard

With all the data collected, DevLake's DORA dashboard is ready to deliver your DORA metrics and benchmarks. You can find the DORA dashboard within the Grafana instance shipped with DevLake, ready for you to put into action.

You can customize the DORA dashboard by editing the underlying SQL query of each panel.

For a breakdown of each metric's SQL query, please refer to the corresponding metric docs:

If you aren't familiar with Grafana, please refer to our Grafana doc, or jump into Slack for help.


🎉🎉🎉 Congratulations! You are now a DevOps Hero, with your own DORA dashboard!