We have developed several worker for the [Security Inventory beta](https://gitlab.com/groups/gitlab-org/-/epics/16484) and there are a few performance tweaks we should consider. ## Set defer_on_database_health_signal This is likely a good place to start. Should we implement this for all of our workers? Should we do it gradually, or for all of them at once? Does it make sense to use the same `delay_by` value for each of them? * Implementation: https://gitlab.com/gitlab-org/gitlab/-/blob/master/app/workers/concerns/worker_attributes.rb#L249 * Example usage: https://gitlab.com/gitlab-org/gitlab/-/blob/master/ee/app/workers/sbom/build_dependency_graph_worker.rb * Example usage: https://gitlab.com/gitlab-org/gitlab/-/blob/master/ee/app/workers/security/generate_policy_violation_comment_worker.rb * Example usage: https://gitlab.com/gitlab-org/gitlab/-/blob/master/ee/app/workers/search/zoekt/indexing_task_worker.rb ## Set concurrency_limit This may not be as relevant, but worth considering. Paraphrasing @ghavenga: "The concurrency limit is not reactive to system growth, so whatever limit you set might not scale well in the future." * Implementation: https://gitlab.com/gitlab-org/gitlab/-/blob/master/app/workers/concerns/worker_attributes.rb * Example usage: https://gitlab.com/gitlab-org/gitlab/-/blob/master/ee/app/workers/security/sync_project_policy_worker.rb * Example usage: https://gitlab.com/gitlab-org/gitlab/-/blob/master/ee/app/workers/search/zoekt/indexing_task_worker.rb