codspeed workflow (pydantic/pydantic)

The codspeed workflow from pydantic/pydantic, explained and optimized by Latchkey.

CI health: D - needs work

Point runs-on at Latchkey and get caching, run de-duplication, job timeouts, self-healing for flaky steps, and up to 58% lower cost, applied automatically.

Grade your own workflow free or run it on Latchkey →

Source: pydantic/pydantic.github/workflows/codspeed.ymlLicense MITView source

What it does

This is the codspeed workflow from the pydantic/pydantic repository, a real project running GitHub Actions. It is shown here with attribution under its MIT license.

Below, Latchkey shows a faster, safer version produced by its optimization engine.

The workflow

workflow (.yml)

name: codspeed

on:
  push:
    branches:
      - main
  pull_request:
  # `workflow_dispatch` allows CodSpeed to trigger backtest
  # performance analysis in order to generate initial data.
  workflow_dispatch:

env:
  UV_FROZEN: true
  UV_PYTHON: 3.14

permissions:
  contents: read

jobs:
  codspeed-profiling:
    name: CodSpeed profiling
    runs-on: ubuntu-24.04
    steps:
      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
        with:
          persist-credentials: false

      - uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # v8.2.0

      # Using this action is still necessary for CodSpeed to work:
      - uses: actions/setup-python@ece7cb06caefa5fff74198d8649806c4678c61a1 # v6.3.0
        with:
          python-version: ${{ env.UV_PYTHON }}

      - name: install deps
        run: uv sync --all-packages --group testing-extra --extra email --frozen

      - name: install rust stable
        # no release versioning, see https://github.com/dtolnay/rust-toolchain/issues/180
        uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88
        with:
          components: llvm-tools
          toolchain: stable

      - name: cache rust
        uses: Swatinem/rust-cache@c19371144df3bb44fab255c43d04cbc2ab54d1c4 # v2.9.1
        with:
          workspaces: pydantic-core

      - name: install pydantic-core with profiling symbols
        run: |
          cd pydantic-core
          uv pip uninstall pytest-speed
          make build-pgo
        env:
          CARGO_PROFILE_RELEASE_DEBUG: 'line-tables-only'
          CARGO_PROFILE_RELEASE_STRIP: 'false'

      - name: Run CodSpeed benchmarks
        uses: CodSpeedHQ/action@a4a36bb07c0638b0b4ca52bf1f3dad1b4289e52f # v4.18.1
        with:
          mode: simulation
          run: uv run --no-sync pytest ./tests/benchmarks ./tests/pydantic_core/benchmarks --codspeed

The same workflow, on Latchkey

Estimated ~20% faster on cache hits, plus fewer wasted runs and a safer supply chain. Added and changed lines are highlighted.

name: codspeed
 
on:
  push:
    branches:
      - main
  pull_request:
  # `workflow_dispatch` allows CodSpeed to trigger backtest
  # performance analysis in order to generate initial data.
  workflow_dispatch:
 
env:
  UV_FROZEN: true
  UV_PYTHON: 3.14
 
permissions:
  contents: read
 
concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true
 
jobs:
  codspeed-profiling:
    timeout-minutes: 30
    name: CodSpeed profiling
    runs-on: latchkey-small
    steps:
      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
        with:
          persist-credentials: false
 
      - uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # v8.2.0
 
      # Using this action is still necessary for CodSpeed to work:
      - uses: actions/setup-python@ece7cb06caefa5fff74198d8649806c4678c61a1 # v6.3.0
        with:
          cache: 'pip'
          python-version: ${{ env.UV_PYTHON }}
 
      - name: install deps
        run: uv sync --all-packages --group testing-extra --extra email --frozen
 
      - name: install rust stable
        # no release versioning, see https://github.com/dtolnay/rust-toolchain/issues/180
        uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88
        with:
          components: llvm-tools
          toolchain: stable
 
      - name: cache rust
        uses: Swatinem/rust-cache@c19371144df3bb44fab255c43d04cbc2ab54d1c4 # v2.9.1
        with:
          workspaces: pydantic-core
 
      - name: install pydantic-core with profiling symbols
        run: |
          cd pydantic-core
          uv pip uninstall pytest-speed
          make build-pgo
        env:
          CARGO_PROFILE_RELEASE_DEBUG: 'line-tables-only'
          CARGO_PROFILE_RELEASE_STRIP: 'false'
 
      - name: Run CodSpeed benchmarks
        uses: CodSpeedHQ/action@a4a36bb07c0638b0b4ca52bf1f3dad1b4289e52f # v4.18.1
        with:
          mode: simulation
          run: uv run --no-sync pytest ./tests/benchmarks ./tests/pydantic_core/benchmarks --codspeed

What changed

Run on Latchkey managed runners with one line (runs-on), which apply the fixes below automatically and self-heal transient failures. This example uses latchkey-small; pick the runner size that fits the job.
Cancel superseded runs when a branch or PR gets a newer push.
Cache dependency installs on the setup step so they are served from cache.
Add a job timeout so a hung step cannot burn hours of runner time.

This workflow runs 1 job per trigger. On Latchkey the same minutes cost up to 58% less than GitHub-hosted, with zero queue time.

Actions used in this workflow

actions/checkout astral-sh/setup-uv actions/setup-python Swatinem/rust-cache