All updates
QAFrontend

Internal Fix: Repair QA Release-Notes Automation Pipeline

PR #820heytulsiprasadMar 3, 2026 · 20:14 UTC
QAMar 3, 2026

Executive summary

An internal continuous-integration fix that repairs the automated QA release-notes pipeline so it correctly packages its working files and routes generated entries to the right destination repository. This is tooling-only and has no effect on any customer-facing DSM application.

Why this was needed

The QA release-summarize workflow was failing to deliver release-note entries for two reasons. First, the build artifact that carries the pipeline's working data (.release_notes_work) is a hidden (dot-prefixed) directory, and the upload step was silently skipping its contents. Second, the workflow dispatched its results to a repository name with incorrect casing (dsm-release-notes), which did not match the actual target repository, so the dispatch never reached its intended destination.

Client / user impact

No impact on end users, clients, or any DSM product surface (Mail, Admin, Vendor, SSO). The benefit is internal: the QA release-notes automation can once again capture changes and publish summary entries reliably, restoring the changelog pipeline the team uses to track what ships to QA.

Technical scope

  • .github/workflows/qa-release-summarize.yml: added include-hidden-files: true to the upload-artifact step so the hidden .release_notes_work directory's contents are actually uploaded; corrected the --dispatch-repo argument to Datagain-Services/DSM_Release_Notes.
  • scripts/release_notes/finalize_and_dispatch.py: updated the DISPATCH_REPO_DEFAULT constant to Datagain-Services/DSM_Release_Notes.
  • scripts/release_notes/README.md: documentation updated to reference the corrected target repo and the dispatch-token requirement against it.
  • Net change is tiny (3 files, +5 / -4); no application or runtime code touched.

Risk & mitigation

Very low. Changes are confined to CI workflow configuration, an automation helper script, and its docs; no product code, API, or database is affected. The only meaningful risk is the dispatch target casing being wrong again, which is mitigated by the same value now living in a single Python default constant and being referenced consistently in the workflow and README.

QA validation focus

  • Trigger the QA release-summarize workflow and confirm the release-notes-work artifact is uploaded with the hidden .release_notes_work contents present (not empty).
  • Confirm a qa_release_entry.v1 repository-dispatch event is received by Datagain-Services/DSM_Release_Notes.
  • Verify the dispatch token has permission to call repository dispatch on the corrected repo name.
  • No regression testing of DSM end-user applications is required.