This GitHub Action provides a way to evaluate statistics for your GitHub Actions workflows. With this action, you can easily monitor the performance of your workflows and identify areas for improvement.
Metrics that are evaluated are:
| Metric | Unit of measure | Description |
|---|---|---|
| Average duration | second | Average of the successful workflow runs, with conclusion status of either successful or skipped. |
| Median duration | second | Median of the successful workflow runs, with conclusion status of either successful or skipped. |
| Total number of runs | workflow run | Total number of workflow runs. |
| Success rate | percentage | Percentage of successful runs for the workflow, with conclusion status of either successful or skipped. |
- As a product engineer, I want to identify areas of improvement for my process automation, so that I can improve delivery in next iteration.
- As an engineering manager, I want to identify waste and inefficiencies in my SDLC process, so that I can reduce cycle time and improve velocity.
- As a DevOps platform owner, I want to identify long running workflows, so that I can right-sizing the runners.
The following options are available for configuring the action:
| Configuration | Required | Default | Description |
|---|---|---|---|
GH_TOKEN |
Yes | N/A | A GitHub token with access to the repository. Minimal scope is repo |
OWNER_NAME |
Yes | N/A | Name of the repository owner. |
REPO_NAME |
No | N/A | Name of the repository. If REPO_NAME is not provided, the action will analyse all the workflow runs in the organisation. |
START_DATE |
Yes | N/A | Start date for the workflow runs data set. This should be in the format YYYY-MM-DD. |
END_DATE |
Yes | N/A | End date for the workflow runs data set. This should be in the format YYYY-MM-DD. |
DELAY_BETWEEN_QUERY |
No | N/A | No. of seconds to wait between queries to the GitHub API. This is to prevent errors from rate limiting when analysing the whole org. |
workflow-names.txt |
No | N/A | A file that contains a list of selected workflow names to filter the result. This should be in the runner's workspace folder. |
After the action has completed, two files will be created in the root of the runner workspace:
runs.jsonororg-runs.json- a JSON array of all workflow runs in the specified time range for the specified repository or organization.workflow-stats.csvororg-workflow-stats.csv- a CSV file with workflow run statistics for the specified repository or organization.
These are data files that then can be used for further analysis or reporting in visualizer of your choice. For example, you can ingest into datastore and visualize with PowerBI. Below are some examples on generating markdown table and mermaid diagram with the data files
To use this action, simply include it in your workflow file:
This will analyse workflow runs in the selected repository, including the durations and success rate of each workflow.
name: My Workflow
on: workflow_dispatch
jobs:
evaluate-actions-consumption:
runs-on: ubuntu-latest
steps:
- name: Call workflow-runs action
uses: kittychiu/[email protected]
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
OWNER_NAME: "myOrg"
REPO_NAME: "myRepo"
START_DATE: "2023-07-01"
END_DATE: "2023-08-01"
- name: Upload all .txt .csv .md files to artifact
uses: actions/upload-artifact@v3
with:
name: workflow-stats
path: |
workflow-stats.csv
runs.jsonBelow is an example of the workflow-stats.csv file:
workflow_name,average_duration,median_duration,success_rate,total_runs
workflow_1,12.33,12.00,100.00,3
workflow_3,25.12,22.00,20.93,43
workflow_2,15.50,15.50,50.00,2
This will further convert workflow-stats.csv file containing workflow metrics into a markdown table, mermaid diagram, and publishes it to a new issue. An example of the rendered outputs is in this Issue.
name: Weekly Retrospective Report
on:
schedule:
- cron: '0 12 * * 5'
jobs:
evaluate-actions-consumption:
runs-on: ubuntu-latest
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
OWNER_NAME: ${{ github.repository_owner }}
steps:
- name: Checkout workflow-names.txt
uses: actions/checkout@v3
- name: Set dates and repo name
run: |
echo "START_DATE=$(date -d '-1 month' +%Y-%m-%d)" >> "$GITHUB_ENV"
echo "END_DATE=$(date +%Y-%m-%d)" >> "$GITHUB_ENV"
repo=$(echo "${{ github.repository }}" | cut -d'/' -f2)
echo "REPO_NAME=${repo}" >> $GITHUB_ENV
- name: Call workflow-runs action
uses: kittychiu/[email protected]
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
REPO_NAME: ${{ env.REPO_NAME }}
START_DATE: ${{ env.START_DATE }}
END_DATE: ${{ env.END_DATE }}
- name: Convert workflow-stats.CSV to stats-table.md markdown table
run: |
echo -e "## Table View\n" > stats-table.md
header=$(head -n 1 workflow-stats.csv | sed 's/,/|/g' | sed 's/_/ /g')
echo -e "|${header}|" >> stats-table.md
metadata=$(head -n 1 workflow-stats.csv | sed 's/,/|/g' | sed 's/[^|]/-/g')
echo -e "|${metadata}|" >> stats-table.md
tail -n +2 workflow-stats.csv | sed 's/,/|/g; s/^/|/; s/$/|/' >> stats-table.md
- name: Convert workflow-stats.CSV to stream-diagram.md mermaid diagram
run: |
echo -e "## Value Stream View\n" > stream-diagram.md
echo -e '```mermaid' >> stream-diagram.md
echo -e 'timeline' >> stream-diagram.md
head -n 1 workflow-stats.csv | sed 's/,/ : /g' | sed 's/_/ /g' | awk -F'|' '{for(i=1;i<=NF;i++) printf("%s%s", " ", $i, i==NF?"\n":", ")}' | sed 's/^/ /' >> stream-diagram.md
tail -n +2 workflow-stats.csv | sed 's/,/ : /g' | awk -F'|' '{for(i=1;i<=NF;i++) printf("%s%s", "\n ", $i, i==NF?"\n":", ")}' | sed 's/^/ /' >> stream-diagram.md
echo -e '\n```' >> stream-diagram.md
- name: Combine into issue content
run: |
echo "Combine output files"
cat stream-diagram.md stats-table.md > issue_view.md
- name: Publish content to a new GitHub Issue
uses: peter-evans/create-issue-from-file@v4
with:
title: Workflow runs summary `${{ env.REPO_NAME }}` repo (${{ env.START_DATE }} - ${{ env.END_DATE }})
content-filepath: issue_view.md
- name: Upload all .txt .csv .md files to artifact
uses: actions/upload-artifact@v3
with:
name: workflow-stats
path: |
stats-table.md
stream-diagram.md
workflow-stats.csv
runs.jsonExample content of workflow-names.txt:
workflow_1
workflow_2
workflow_3
Below is an example of the stats-table.md file:
|workflow name|average duration|median duration|success rate|total runs|
|-------------|----------------|---------------|------------|----------|
|workflow_1|17.00|17.00|100.00|1|
|workflow_2|36.17|36.50|53.70|54|
|workflow_3|3.00|2.00|100.00|3|Below is an example of the stream-diagram.md file:
timeline
workflow name : average duration : median duration : success rate : total runs
CI Build : 17.00 : 17.00 : 100.00 : 1
QA & Validation : 36.17 : 36.50 : 53.70 : 54
Deploy to non-prod : 3.00 : 2.00 : 100.00 : 3
This will analyse workflow runs in the selected organisation, including all workflows for each repository. An example of the rendered output is in this Issue.
name: Monthly SLOs Report
on:
schedule:
- cron: '0 0 1 * *'
jobs:
evaluate-actions-consumption:
runs-on: ubuntu-latest
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
steps:
- name: Set dates
run: |
echo "START_DATE=$(date -d '-14 days' +%Y-%m-%d)" >> "$GITHUB_ENV"
echo "END_DATE=$(date +%Y-%m-%d)" >> "$GITHUB_ENV"
- name: Test docker action
uses: kittychiu/[email protected]
env:
GH_TOKEN: ${{ env.GH_TOKEN }}
OWNER_NAME: ${{ github.repository_owner }}
START_DATE: ${{ env.START_DATE }}
END_DATE: ${{ env.END_DATE }}
DELAY_BETWEEN_QUERY: 5
- name: Convert org-workflow-stats.csv to stats-table.md markdown table
run: |
echo -e "## Table View\n" > stats-table.md
header=$(head -n 1 org-workflow-stats.csv | sed 's/,/|/g' | sed 's/_/ /g')
echo -e "|${header}|" >> stats-table.md
metadata=$(head -n 1 org-workflow-stats.csv | sed 's/,/|/g' | sed 's/[^|]/-/g')
echo -e "|${metadata}|" >> stats-table.md
tail -n +2 org-workflow-stats.csv | sed 's/,/|/g; s/^/|/; s/$/|/' >> stats-table.md
- name: Publish result to a new issue
uses: peter-evans/create-issue-from-file@v4
with:
title: Workflow runs summary for `${{ env.OWNER_NAME }}` org (${{ env.START_DATE }} - ${{ env.END_DATE }})
content-filepath: stats-table.md
- name: Upload all .txt .csv .md files to artifact
uses: actions/upload-artifact@v3
with:
name: workflow-stats
path: |
stats-table.md
org-workflow-stats.csv
org-runs.jsonBelow is an example of the stats-table.md file:
|repository name|workflow name|average duration|median duration|success rate|total runs|
|---------------|-------------|----------------|---------------|------------|----------|
|repo_1|Test|3.00|3.00|100.00|1|
|repo_1|Build|20.20|17.00|80.00|5|
|repo_1|Deploy|17.00|17.00|100.00|1|
|repo_2|Custom Validation|2.00|2.00|100.00|1|
|repo_2|Linter|2.00|2.00|100.00|1|
|repo_3|Superlinter|25.38|23.00|30.00|50|
|repo_3|Long Build|36.17|36.50|53.70|54|
|repo_3|Smoke Test|19.69|14.00|23.08|13|Please see the contributing guidelines for more information.
This project is licensed under the MIT License.