Our Tekton instance is accumulating repositories, artifacts, jobs, and users. Which metrics and components should I monitor first, and what usually helps most with database performance, storage, background jobs, CI workload isolation, caching, and horizontal scaling?