[internal] Produce data for external analysis #24119
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
name: "[internal] Produce data for external analysis" | |
on: | |
workflow_call: | |
workflow_dispatch: | |
inputs: | |
test_workflow_run_id: | |
description: "Unique GitHub workflow run ID to use for data" | |
default: 10066309412 | |
type: number | |
test_workflow_run_attempt: | |
description: "Run attempt of the workflow run" | |
default: 1 | |
type: number | |
upload_data: | |
description: "Upload data to datastore cluster for our dashboard" | |
default: false | |
type: boolean | |
workflow_run: | |
workflows: | |
- "All post-commit tests" | |
- "(Single-card) Model perf tests" | |
- "(Single-card) Device perf tests" | |
- "(Single-card) Demo tests" | |
- "Nightly fast dispatch tests" | |
- "(T3K) T3000 demo tests" | |
- "(T3K) T3000 model perf tests" | |
- "(T3K) T3000 perplexity tests" | |
- "(T3K) T3000 model perf tests" | |
- "(T3K) T3000 profiler tests" | |
- "(T3K) T3000 unit tests" | |
- "(TG) TG unit tests" | |
- "(TG) TG demo tests" | |
- "(TG) TG frequent tests" | |
- "(TG) TG model perf tests" | |
- "(TGG) TGG model perf tests" | |
- "(TGG) TGG unit tests" | |
- "(TGG) TGG demo tests" | |
- "(TGG) TGG frequent tests" | |
- "ttnn - Run sweeps" | |
- "Blackhole post-commit tests" | |
types: | |
- completed | |
jobs: | |
produce-cicd-data: | |
runs-on: ubuntu-latest | |
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" >> "$GITHUB_OUTPUT" | |
echo "::notice title=target-workflow-link::The workflow being analyzed is available at https://github.com/tenstorrent/tt-metal/actions/runs/$run_id/attempts/$attempt_number" | |
- name: Output auxiliary values | |
env: | |
GH_TOKEN: ${{ github.token }} | |
run: | | |
echo "[Info] Workflow run attempt" | |
gh api /repos/tenstorrent/tt-metal/actions/runs/${{ steps.get-run-id-and-attempt.outputs.run-id }}/attempts/${{ steps.get-run-id-and-attempt.outputs.attempt-number }} | |
gh api /repos/tenstorrent/tt-metal/actions/runs/${{ steps.get-run-id-and-attempt.outputs.run-id }}/attempts/${{ steps.get-run-id-and-attempt.outputs.attempt-number }} > workflow.json | |
echo "[Info] Workflow run attempt jobs" | |
gh api /repos/tenstorrent/tt-metal/actions/runs/${{ steps.get-run-id-and-attempt.outputs.run-id }}/attempts/${{ steps.get-run-id-and-attempt.outputs.attempt-number }}/jobs --paginate | |
# GitHub chunks the jobs array into multiple objects side-by-side if there are more than 100 jobs, so we need to slurp them into one giant object for later Python processing | |
gh api /repos/tenstorrent/tt-metal/actions/runs/${{ steps.get-run-id-and-attempt.outputs.run-id }}/attempts/${{ steps.get-run-id-and-attempt.outputs.attempt-number }}/jobs --paginate | jq -s '{total_count: .[0].total_count, jobs: map(.jobs) | add}' > workflow_jobs.json | |
- name: Collect workflow artifact and job logs | |
shell: bash | |
env: | |
GH_TOKEN: ${{ github.token }} | |
run: | | |
./infra/data_collection/github/download_cicd_logs_and_artifacts.sh --workflow-run-id ${{ steps.get-run-id-and-attempt.outputs.run-id }} --attempt-number ${{ steps.get-run-id-and-attempt.outputs.attempt-number }} | |
find generated/cicd/ -type f | |
- uses: actions/setup-python@v5 | |
with: | |
python-version: '3.8' | |
cache: 'pip' | |
cache-dependency-path: 'infra/requirements-infra.txt' | |
- name: Install infra dependencies | |
run: pip install -r infra/requirements-infra.txt | |
- name: Create JSON | |
env: | |
PYTHONPATH: ${{ github.workspace }} | |
run: python3 .github/scripts/data_analysis/create_pipeline_json.py | |
- name: Show directory to see output files | |
run: ls -hal | |
- name: Upload cicd data | |
uses: ./.github/actions/upload-data-via-sftp | |
if: ${{ github.event_name == 'workflow_run' || inputs.upload_data }} | |
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_*.json | |
workflow.json | |
workflow_jobs.json | |
- uses: ./.github/actions/slack-report | |
if: ${{ failure() }} | |
with: | |
slack_webhook_url: ${{ secrets.SLACK_WEBHOOK_URL }} | |
owner: U014XCQ9CF8 # tt-rkim | |
test-produce-benchmark-environment: | |
runs-on: ubuntu-latest | |
steps: | |
- uses: actions/checkout@v4 | |
- name: create dummy | |
shell: bash | |
run: | | |
mkdir -p generated/benchmark_data | |
touch 'generated/benchmark_data/measurement_2024-07-12T05:01:45+0000.csv' | |
touch 'generated/benchmark_data/measurement_2024-07-12T04:59:14+0000.csv' | |
touch 'generated/benchmark_data/measurement_2024-07-12T05:03:29+0000.csv' | |
- uses: actions/[email protected] | |
with: | |
python-version: '3.10' | |
cache: 'pip' | |
cache-dependency-path: 'infra/requirements-infra.txt' | |
- name: Install infra dependencies | |
run: pip install -r infra/requirements-infra.txt | |
- name: Create environment CSV test | |
env: | |
PYTHONPATH: ${{ github.workspace }} | |
ARCH_NAME: grayskull | |
run: python3 .github/scripts/data_analysis/create_benchmark_environment_csv.py | |
- name: Show files | |
shell: bash | |
run: find generated/benchmark_data -type f | xargs -n 1 -I {} bash -c 'echo {} && cat {}' |