Feature request
The setup-helm action currently only uses the runner's local tool cache (/opt/hostedtoolcache/) which does not persist across different GitHub Actions runners. This means Helm is re-downloaded on every workflow run, even when using the same version.
Unlike other setup actions (e.g., actions/setup-go, actions/setup-node), the Helm installation is not visible in GitHub's cache dashboard and cannot be shared across workflow runs. This results in:
- Slower workflow runs due to repeated downloads
- Increased bandwidth usage
- Inconsistent behavior compared to other GitHub setup actions
Users can currently work around this by manually adding an actions/cache step before setup-helm, but this requires understanding the internal cache directory structure and adds boilerplate to every workflow.
Solution
Integrate @actions/cache directly into the action to enable persistent caching across workflow runs. This would:
- Add
@actions/cache as a dependency
- Restore the cache from GitHub's cache storage before checking the local tool cache
- Save the cache after successful Helm installation
- Make cached Helm installations visible in the repository's cache dashboard
This could be implemented as an opt-in or opt-out feature via a new input parameter:
- uses: azure/setup-helm@v4
with:
version: 'v3.18.4'
cache: true # Enable persistent caching (could default to true)
The cache key should include the OS, architecture, and Helm version to ensure correct cache invalidation when any of these change.
Feature request
The
setup-helmaction currently only uses the runner's local tool cache (/opt/hostedtoolcache/) which does not persist across different GitHub Actions runners. This means Helm is re-downloaded on every workflow run, even when using the same version.Unlike other setup actions (e.g.,
actions/setup-go,actions/setup-node), the Helm installation is not visible in GitHub's cache dashboard and cannot be shared across workflow runs. This results in:Users can currently work around this by manually adding an
actions/cachestep beforesetup-helm, but this requires understanding the internal cache directory structure and adds boilerplate to every workflow.Solution
Integrate
@actions/cachedirectly into the action to enable persistent caching across workflow runs. This would:@actions/cacheas a dependencyThis could be implemented as an opt-in or opt-out feature via a new input parameter:
The cache key should include the OS, architecture, and Helm version to ensure correct cache invalidation when any of these change.