M5: Patch Deployment & Job Management

Backend:
- migrations/003_jobs_scheduling.sql: retry_next_at/last_error columns,
  pg_notify trigger for immediate job dispatch, retry index
- pm-agent-client: ApplyPatchesRequest/Response, AgentJobStatus,
  RollbackResponse types; apply_patches/job_status/rollback_job
  client methods + generic POST helper
- pm-core/models: JobStatus, JobKind, PatchJob, PatchJobHost,
  CreateJobRequest, PatchJobSummary
- pm-web/routes/jobs.rs: POST/GET /api/v1/jobs, GET /jobs/:id,
  POST /jobs/:id/cancel, POST /jobs/:id/rollback
- pm-worker/job_executor.rs: NOTIFY listener, periodic scanner,
  execute_host_job, poll_running_jobs, handle_host_failure (3-retry
  exponential backoff 1m/5m/30m), sync_job_status, retry_pending_jobs
- pm-worker/main.rs: spawn job_executor

Frontend:
- types/index.ts: PatchInfo, PatchJobHost, PatchJob, PatchJobSummary,
  CreateJobRequest interfaces
- api/client.ts: jobsApi (list/get/create/cancel/rollback),
  patchesApi (getHostPatches)
- pages/PatchDeploymentPage.tsx: 3-step MUI Stepper
  (host select → configure → result)
- pages/JobsPage.tsx: job list table, expandable per-host detail,
  cancel/rollback actions with confirm dialog, load-more pagination
- App.tsx: /jobs and /deployment routes wired to real pages

cargo check: 0 errors | vite build: 0 errors

migrations/003_jobs_scheduling.sql

@@ -0,0 +1,44 @@
+-- Migration: 003_jobs_scheduling
+-- Description: Retry/scheduling columns for patch_job_hosts, and NOTIFY trigger
+--              for immediate patch job dispatch.
+
+-- ============================================================
+-- Add retry-scheduling columns to patch_job_hosts
+-- ============================================================
+
+-- When the retry engine should next attempt this host; NULL = not scheduled
+ALTER TABLE patch_job_hosts
+    ADD COLUMN retry_next_at TIMESTAMPTZ;
+
+-- Last failure reason captured by the worker for display in the UI
+ALTER TABLE patch_job_hosts
+    ADD COLUMN last_error TEXT;
+
+-- ============================================================
+-- pg_notify trigger: fires when an immediate job is inserted
+-- ============================================================
+
+CREATE OR REPLACE FUNCTION notify_job_enqueued()
+    RETURNS TRIGGER
+    LANGUAGE plpgsql
+AS $$
+BEGIN
+    IF NEW.immediate = TRUE THEN
+        PERFORM pg_notify('job_enqueued', NEW.id::text);
+    END IF;
+    RETURN NEW;
+END;
+$$;
+
+CREATE TRIGGER trg_job_enqueued
+    AFTER INSERT ON patch_jobs
+    FOR EACH ROW
+    EXECUTE FUNCTION notify_job_enqueued();
+
+-- ============================================================
+-- Index: efficiently find hosts due for retry
+-- ============================================================
+
+CREATE INDEX idx_pjh_retry
+    ON patch_job_hosts (retry_next_at)
+    WHERE retry_next_at IS NOT NULL;