CI Scheduled Workflow Health

Scheduled Workflow Health Report

Generates an HTML report showing the health status of all scheduled GitHub Actions workflows in the camunda/camunda repository.

Then summarizes the concerning failing and flaky workflows by owner.

What it does

1. Lists all workflow files from git show main:.github/workflows/
2. Filters to those containing a schedule: trigger
3. Extracts the # owner: comment from the first 20 lines of each YAML file
4. Queries the GitHub API (via gh) for the last 5 scheduled runs on main
5. Classifies each workflow:
   • failing — most recent run failed AND all checked runs failed
   • flaky — most recent run failed BUT some checked runs passed
   • ok — most recent run passed
   • no_runs / in_progress — no completed scheduled runs found
6. Outputs scheduled-workflow-report.html with collapsible sections, owner info, and a legend

Prerequisites

• git with access to the main branch (fetched)
• gh CLI authenticated (gh auth status)
• python3

How to run

python3 .claude/skills/ci-scheduled-workflow-health/scripts/scheduled-workflow-report.py

The script is at: .claude/skills/ci-scheduled-workflow-health/scripts/scheduled-workflow-report.py.

Output: scheduled-workflow-report.html in the current directory. Open with a browser.

stderr shows progress and a summary of failing/flaky workflows with their owners.

Customization

• RUNS_TO_CHECK constant (default 5) controls how many recent runs to inspect per workflow
• OWNER / REPO constants control the target repository
• Owner extraction looks for # owner: <team> in the first 20 lines of each workflow YAML

Key design decisions

• Uses git show main: instead of filesystem reads to always reflect the main branch state
• Uses gh api (no pagination) since we only need ≤5 runs per workflow
• Filters API calls to event=schedule&branch=main to only see scheduled runs
• HTML report uses <details> for collapsible sections — failing/flaky are open by default