Health check endpoint.

Reports status: "active" while any HTTP request OR webhook background task is in flight, so the K8s preStop drain loop waits for both kinds of work before SIGTERM is delivered. Without the webhook coupling, preStop would see a quiet HTTP queue and exit early — burning the grace period on a pod that still has accepted (and 200'd) webhook work pending; the in-lifespan drain would then have far less than the chart's nominal terminationGracePeriodSeconds to finish that work.