[AUTO-INF-1] Avoid reinstalling Helm and Kubeconform per CI job #10363

Open
opened 2026-04-18 09:11:23 +00:00 by HAL9000 · 0 comments
Owner

Metadata

  • Commit message: infra: cache helm and kubeconform binaries across CI jobs
  • Branch name: infra/cache-helm-kubeconform-ci

Background and Context

Currently, Helm and Kubeconform are downloaded and installed separately in the unit_tests, integration_tests, and e2e_tests nox environments, adding ~2 min per install. This repeated installation inflates overall pipeline setup time on every CI run, slowing down developer feedback loops and consuming unnecessary compute resources.

Expected Behavior

Helm and Kubeconform binaries are downloaded once per CI pipeline run and cached, then reused across all nox environments (unit_tests, integration_tests, e2e_tests) without re-downloading or reinstalling.

Acceptance Criteria

  • Helm binary is downloaded once and cached at the pipeline level (e.g., via CI cache key on version + OS)
  • Kubeconform binary is downloaded once and cached at the pipeline level
  • All three nox environments (unit_tests, integration_tests, e2e_tests) consume the cached binaries without re-downloading
  • CI pipeline setup time is reduced by ~2 min per job that previously installed these tools
  • Cache invalidation occurs correctly when the pinned version of Helm or Kubeconform changes
  • No functional regression in tests that depend on Helm or Kubeconform

Subtasks

  • Identify current install steps for Helm and Kubeconform in each nox session
  • Define a shared CI cache key strategy (tool name + version + OS/arch)
  • Implement a single download-and-cache step early in the pipeline
  • Update unit_tests, integration_tests, and e2e_tests nox sessions to reference the cached binary path
  • Validate cache hit/miss behaviour in CI (dry run or test pipeline)
  • Update CI documentation / comments to reflect the new caching approach

Definition of Done

This issue should be closed when:

  • Helm and Kubeconform are installed exactly once per pipeline run and reused from cache in all affected nox environments
  • CI pipeline wall-clock time is measurably reduced (target: ~2 min saved per job)
  • All existing tests continue to pass
  • The caching strategy is documented in CI configuration comments

Duplicate Check

Searched open issues pages 1-3 and closed issues page 1 with keywords: helm install, CI speed, job caching. No existing issues found.

Proposal

Currently, Helm and Kubeconform are downloaded and installed separately in the unit_tests, integration_tests, and e2e_tests nox environments, adding ~2 min per install. We should download and cache these binaries once and share them across CI jobs to reduce overall pipeline setup time.


Automated by CleverAgents Bot
Supervisor: Test Infra Pool | Agent: test-infra-pool-supervisor

## Metadata - **Commit message:** `infra: cache helm and kubeconform binaries across CI jobs` - **Branch name:** `infra/cache-helm-kubeconform-ci` ## Background and Context Currently, Helm and Kubeconform are downloaded and installed separately in the `unit_tests`, `integration_tests`, and `e2e_tests` nox environments, adding ~2 min per install. This repeated installation inflates overall pipeline setup time on every CI run, slowing down developer feedback loops and consuming unnecessary compute resources. ## Expected Behavior Helm and Kubeconform binaries are downloaded once per CI pipeline run and cached, then reused across all nox environments (`unit_tests`, `integration_tests`, `e2e_tests`) without re-downloading or reinstalling. ## Acceptance Criteria - [ ] Helm binary is downloaded once and cached at the pipeline level (e.g., via CI cache key on version + OS) - [ ] Kubeconform binary is downloaded once and cached at the pipeline level - [ ] All three nox environments (`unit_tests`, `integration_tests`, `e2e_tests`) consume the cached binaries without re-downloading - [ ] CI pipeline setup time is reduced by ~2 min per job that previously installed these tools - [ ] Cache invalidation occurs correctly when the pinned version of Helm or Kubeconform changes - [ ] No functional regression in tests that depend on Helm or Kubeconform ## Subtasks - [ ] Identify current install steps for Helm and Kubeconform in each nox session - [ ] Define a shared CI cache key strategy (tool name + version + OS/arch) - [ ] Implement a single download-and-cache step early in the pipeline - [ ] Update `unit_tests`, `integration_tests`, and `e2e_tests` nox sessions to reference the cached binary path - [ ] Validate cache hit/miss behaviour in CI (dry run or test pipeline) - [ ] Update CI documentation / comments to reflect the new caching approach ## Definition of Done This issue should be closed when: - Helm and Kubeconform are installed exactly once per pipeline run and reused from cache in all affected nox environments - CI pipeline wall-clock time is measurably reduced (target: ~2 min saved per job) - All existing tests continue to pass - The caching strategy is documented in CI configuration comments ## Duplicate Check Searched open issues pages 1-3 and closed issues page 1 with keywords: helm install, CI speed, job caching. No existing issues found. ## Proposal Currently, Helm and Kubeconform are downloaded and installed separately in the unit_tests, integration_tests, and e2e_tests nox environments, adding ~2 min per install. We should download and cache these binaries once and share them across CI jobs to reduce overall pipeline setup time. --- **Automated by CleverAgents Bot** Supervisor: Test Infra Pool | Agent: test-infra-pool-supervisor
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
cleveragents/cleveragents-core#10363
No description provided.