Run Benchmarks #503

	name: Run Benchmarks

	on:
	schedule:
	- cron: '0 0 * * *' # Runs daily at midnight
	pull_request:

	jobs:
	benchmark:
	runs-on: ubuntu-latest

	steps:
	- uses: actions/checkout@v4
	- name: fetch submodules
	run: git submodule update --init --recursive
	- uses: actions-rust-lang/setup-rust-toolchain@v1
	with:
	rustflags: '' #Disable. By default this action sets environment variable is set to -D warnings. We manage this in the Makefile
	- name: Setup build env
	run: \|
	os=$(echo "$RUNNER_OS" \| tr '[:upper:]' '[:lower:]')
	./scripts/setup-$os.sh
	shell: bash
	- name: Run Benchmarks
	run: \|
	set -o pipefail
	cargo bench -p containerd-shim-benchmarks -- --output-format bencher \| tee output.txt
	# Download previous benchmark result from cache (if exists)
	- name: Cache benchmark data
	uses: actions/cache@v4
	with:
	path: ./cache
	key: ${{ runner.os }}-benchmark
	- name: Store benchmark result
	uses: benchmark-action/[email protected]
	with:
	name: Criterion.rs Benchmark
	tool: 'cargo'
	output-file-path: output.txt
	github-token: ${{ secrets.GITHUB_TOKEN }}
	# my experimental local benchmarking seems to have a 20% margin of error.
	# So I set the alert threshold to 130% of the previous benchmark result.
	# If the current benchmark result is more than 130% of the previous benchmark result, it will fail.
	alert-threshold: '130%'
	fail-on-alert: true
	alert-comment-cc-users: '@runwasi-committers'
	# Enable Job Summary
	summary-always: true
	# Where the previous data file is stored
	external-data-json-path: ./cache/benchmark-data.json

Provide feedback