Lead Time for Changes

A GitHub Action to roughly calculate DORA lead time for changes This is not meant to be an exhaustive calculation, but we are able to approximate fairly close for most of workflows. Why? Our insights indicated that many applications don't need exhaustive DORA analysis - a high level, order of magnitude result is accurate for most workloads.

Current Calculation

Current Limitations

Open questions

Inputs

workflows: required, string, The name of the workflows to process. Multiple workflows can be separated by ,
owner-repo: optional, string, defaults to the repo where the action runs. Can target another owner or org and repo. e.g. 'samsmithnz/DevOpsMetrics', but will require authenication (see below)
default-branch: optional, string, defaults to main
number-of-days: optional, integer, defaults to 30 (days)
commit-counting-method: #optional, defaults to 'last'. Accepts two values, 'last' - to start timing from the last commit of a PR, and 'first' to start timing from the first commit of a PR
pat-token: optional, string, defaults to ''. Can be set with GitHub PAT token. Ensure that Read access to actions and metadata permission is set. This is a secret, never directly add this into the actions workflow, use a secret.
actions-token: optional, string, defaults to ''. Can be set with ${{ secrets.GITHUB_TOKEN }} in the action
app-id: optional, string, defaults to '', application id of the registered GitHub app
app-install-id: optional, string, defaults to '', id of the installed instance of the GitHub app
app-private-key optional, string, defaults to '', private key which has been generated for the installed instance of the GitHub app. Must be provided without leading '-----BEGIN RSA PRIVATE KEY----- ' and trailing ' -----END RSA PRIVATE KEY-----'.

To test the current repo (same as where the action runs)

- uses: samsmithnz/deployment-frequency@main
  with:
    workflows: 'CI'

To test another repo, with all arguments

- name: Test another repo
  uses: samsmithnz/deployment-frequency@main
  with:
    workflows: 'CI/CD'
    owner-repo: 'samsmithnz/DevOpsMetrics'
    default-branch: 'main'
    number-of-days: 30

To use a PAT token to access another (potentially private) repo:

- name: Test elite repo with PAT Token
  uses: samsmithnz/deployment-frequency@main
  with:
    workflows: 'CI/CD'
    owner-repo: 'samsmithnz/SamsFeatureFlags'
    pat-token: "${{ secrets.PATTOKEN }}"

Use the built in Actions GitHub Token to retrieve the metrix

- name: Test this repo with GitHub Token
  uses: samsmithnz/deployment-frequency@main
  with:
    workflows: 'CI'
    actions-token: "${{ secrets.GITHUB_TOKEN }}"

Gather the metric from another repository using GitHub App authentication method:

- name: Test another repo with GitHub App
  uses: samsmithnz/deployment-frequency@main
  with:
    workflows: 'CI'
    owner-repo: 'samsmithnz/some-other-repo'
    app-id: "${{ secrets.APPID }}"
    app-install-id: "${{ secrets.APPINSTALLID }}"
    app-private-key: "${{ secrets.APPPRIVATEKEY }}"

Setup

Permissions: Read access to actions, metadata, and pull requests

Output

Current output shows the inputs, authenication method, rate limit consumption, and then the actual lead time for changes

Owner/Repo: samsmithnz/SamsFeatureFlags
Workflows: Feature Flags CI/CD
Branch: main
Number of days: 30
Commit counting method 'last' being used
Authentication detected: PAT TOKEN
PR average time duration 5.4612962962963
Workflow average time duration 0.30224537037037
Rate limit consumption: 242 / 5000
Lead time for changes average over last 30 days, is 5.76 hours, with a DORA rating of 'Elite'

In the job summary, we show a badge with details:

Definition: For the primary application or service, how long does it take to go from code committed to code successfully running in production.
Results: Lead time for changes is 5.61 hours with a Elite rating, over the last 30 days.
Details:

Repository: samsmithnz/deployment-frequency using main branch
Workflow(s) used: CI

ayodejiayodele / lead-time-for-changes