Skip to content

[internal] Produce analytic data #726

[internal] Produce analytic data

[internal] Produce analytic data #726

Workflow file for this run

name: "[internal] Produce analytic data"
on:
workflow_call:
workflow_dispatch:
inputs:
test_workflow_run_id:
description: "Unique GitHub workflow run ID to use for data"
type: number
test_workflow_run_attempt:
description: "Run attempt of the workflow run"
default: 1
type: number
workflow_run:
workflows: # List workflow that we want to collect data for
- "Build and Test"
- "On PR"
- "On push"
types:
- completed
jobs:
produce-cicd-data:
runs-on: ubuntu-latest
env:
GH_TOKEN: ${{ github.token }}
steps:
- uses: actions/checkout@v4
- name: Output (safe) pipeline values
run: |
echo "pipeline_id (id / run #): ${{ github.run_id }}/${{ github.run_attempt }}"
echo "submissions_ts: "
echo "start_ts: "
echo "end_ts: "
echo "name: ${{ github.workflow }}, but rk recommended name w/out @: ${{ github.workflow_ref }}"
echo "trigger: ${{ github.event_name }}"
echo "sha: ${{ github.sha }}"
echo "(triggering) author/actor: ${{ github.actor }}"
echo "author/actor: ${{ github.triggering_actor }}"
echo "orchestrator: github (Static)"
echo "docker_image: ${{ job.container.image }}"
echo "build duration is post-process"
- name: Get workflow run_id attempt number to analyze
id: get-run-id-and-attempt
shell: bash
run: |
event_name="${{ github.event_name }}"
if [[ "$event_name" == "workflow_dispatch" ]]; then
run_id="${{ inputs.test_workflow_run_id }}"
attempt_number="${{ inputs.test_workflow_run_attempt }}"
elif [[ "$event_name" == "workflow_run" ]]; then
run_id="${{ github.event.workflow_run.id }}"
attempt_number="${{ github.event.workflow_run.run_attempt }}"
[[ -z "$run_id" ]] && { echo "run_id is empty" ; exit 1; }
[[ -z "$attempt_number" ]] && { echo "attempt_number is empty" ; exit 1; }
else
echo "Unknown event name" && exit 1
fi
echo $run_id
echo $attempt_number
echo "run-id=$run_id" >> "$GITHUB_OUTPUT"
echo "attempt-number=$attempt_number" >> "$GITHUB_OUTPUT"
- name: Collect workflow artifact and job logs
shell: bash
run: |
infra/download_workflow_data.sh ${{ github.repository }} ${{ steps.get-run-id-and-attempt.outputs.run-id }} ${{ steps.get-run-id-and-attempt.outputs.attempt-number }}
- uses: actions/setup-python@v5
with:
python-version: '3.8'
cache: 'pip'
cache-dependency-path: 'infra/requirements.txt'
- name: Install infra dependencies
run: pip install -r infra/requirements.txt
- name: Create JSON
env:
PYTHONPATH: ${{ github.workspace }}
run: |
python3 infra/src/generate_data.py --run_id ${{ steps.get-run-id-and-attempt.outputs.run-id }}
# Workaround: Copy file to avoid GH upload filename limitations
cp pipeline_${{ steps.get-run-id-and-attempt.outputs.run-id }}*.json pipelinecopy_${{ steps.get-run-id-and-attempt.outputs.run-id }}.json
- name: Upload cicd data
uses: ./.github/actions/upload-data-via-sftp
if: ${{ github.event_name == 'workflow_run' }}
with:
ssh-private-key: ${{ secrets.SFTP_CICD_WRITER_KEY }}
sftp-batchfile: .github/actions/upload-data-via-sftp/cicd_data_batchfile.txt
username: ${{ secrets.SFTP_CICD_WRITER_USERNAME }}
hostname: ${{ secrets.SFTP_CICD_WRITER_HOSTNAME }}
- name: Upload workflow run data, even on failure
if: ${{ !cancelled() }}
uses: actions/upload-artifact@v4
with:
name: workflow-run-data
path: |
if-no-files-found: warn
path: |
pipelinecopy_${{ steps.get-run-id-and-attempt.outputs.run-id }}.json
generated/cicd/${{ steps.get-run-id-and-attempt.outputs.run-id }}/workflow.json
generated/cicd/${{ steps.get-run-id-and-attempt.outputs.run-id }}/workflow_jobs.json