- 1 Ignore the return code of reference build and don’t do an upload if it fails but don’t turn red

- 1 Expiration on disabled broken benchmarks
  - Don’t disable right away, 2-3 days to respond and then disable
  - 15 days they haven’t worked on it, to remove
- Pruning Disabled tests
  - 3 Statically pruning known bot dimensions in json generation
  - 3 Dynamically generate list of tests to run
    - Can we run this list tests once on a swarming bot and then trigger the rest based on.  
      - Investigate.  Think it is hard given the recipe setup.
- 4 Disabling vs leaving red
  - SoM integration that follows the state of the bug