Skip to content

Different default API test tiers between workflow_call and workflow_dispatch #134

@shawn-hurley

Description

@shawn-hurley

Description

The api_tests_tiers input has different default values depending on how the workflow is triggered across both reusable workflows.

Locations

  • .github/workflows/global-ci.yml
  • .github/workflows/global-ci-bundle.yml

Details

global-ci.yml

  • workflow_call (line 57): default: make test-tier0 test-tier1 test-hub-api
  • workflow_dispatch (line 139): default: make test-tier0 test-tier1

global-ci-bundle.yml

  • workflow_call (line 114): default: make test-tier0 test-tier1 test-hub-api
  • workflow_dispatch (line 198): default: make test-tier0 test-tier1

Impact

Manual workflow dispatches run fewer tests than scheduled/called workflows. This creates inconsistent test coverage:

  • Automated runs test: tier0, tier1, and hub-api
  • Manual runs test: tier0 and tier1 only (missing hub-api tests)

This could lead to:

  • Hub API bugs slipping through manual testing
  • Confusion about expected test coverage
  • False confidence when manually triggering tests

Recommendation

Standardize the default to include all three test tiers across both trigger types:

default: make test-tier0 test-tier1 test-hub-api

Related Files

  • .github/workflows/global-ci.yml:57
  • .github/workflows/global-ci.yml:139
  • .github/workflows/global-ci-bundle.yml:114
  • .github/workflows/global-ci-bundle.yml:198

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions