tune memory usage + fix dashboard + add justfile recipes
- scale API server to 1 replica
- prometheus: 60s scrape, 7d retention
- executive dashboard: fix pods-not-ready query, pie chart for
flow runs, show work_pool in deployments table
- justfile: add status (with top), parameterized logs, prefect
CLI proxy, dashboards reload
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>