Uploaded image for project: 'AI Platform Core Components'
  1. AI Platform Core Components
  2. AIPCC-3564

Midstream versioning directly in RHAIIS collection

    • Icon: Epic Epic
    • Resolution: Done
    • Icon: Undefined Undefined
    • None
    • None
    • Accelerator Enablement
    • None
    • Midstream versioning directly in RHAIIS collection
    • False
    • Hide

      None

      Show
      None
    • False
    • To Do
    • 0% To Do, 0% In Progress, 100% Done

      We propose to delegate the management of frequent vLLM releases to the RHAIIS team (using AIPCC automation), in order to streamline the RHAIIS release process—making it faster, more efficient, and cost-effective.

      This will involve:

      • Defining the target vLLM version directly within the RHAIIS collection: RHAIIS Collection Repository (e.g., v0.9.2.1). At this stage, we expect only the patch version (last digit) to be updated regularly. Updating major or minor versions typically requires corresponding updates to Torch and related dependencies, which is out of scope for this effort.
      • Configuring Renovate to automate version bumps: Renovate will monitor new midstream vLLM tags and automatically generate merge requests in the RHAIIS pipeline. These MRs will be reviewed by the RHAIIS team. This automation will reduce the manual overhead, enable faster integration of updates, and significantly shorten the feedback loop.

              rh-ee-pmattsso Percy Mattsson
              emacchi@redhat.com Emilien Macchi
              Frank's Team
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated:
                Resolved: