Summary

GitHub container registry randomly hitting api rate limit causing intermittent issue when downloading Trivy container CIS image / db and resulting in downstream scans to fail.

Options

1. Skip cache by default and always leverage upstream DB

• As an emergency failure in case of upstream hosting environment flakiness / failures, there won't be caches since they are expired and skipped

2. Option to override default [<run_id>_<attempt_id>] to make it save unique caches in matrix job
   a. specify option to override cache key as input
   b. Detect if running in matrix context and generate unique caches based on some parameter.

• Either way, this is also NOT recommended since multiple cache keys might exhaust Github default cache limits of 10GB since each cache key needs to be unique for run/attempt/matrix-input combination

3. Host a vuln DB mirror as Kong repo and use it as the DB URL override

• Overhead of effort and maintenance to maintain a list of updated offline Trivy vulnerability DB and distributing them