feat: Remove Run Now, add source tracking, optimize dashboard

- Remove /run-now endpoint (use task priority instead)
- Add source tracking to worker_tasks (source, source_schedule_id, source_metadata)
- Parallelize dashboard API calls (Promise.all)
- Add 1-5 min caching to /markets/dashboard and /national/summary
- Add performance indexes for dashboard queries

Migrations:
- 104: Task source tracking columns
- 105: Dashboard performance indexes

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/migrations/104_task_source_tracking.sql

@@ -0,0 +1,25 @@
+-- Migration 104: Add source tracking to worker_tasks
+-- Purpose: Track WHERE tasks are created from (schedule vs API endpoint)
+--
+-- All automated task creation should be visible in task_schedules.
+-- This column helps identify "phantom" tasks created outside the schedule system.
+
+-- Add source column to worker_tasks
+ALTER TABLE worker_tasks
+ADD COLUMN IF NOT EXISTS source VARCHAR(100);
+
+-- Add source_id column (references schedule_id if from a schedule)
+ALTER TABLE worker_tasks
+ADD COLUMN IF NOT EXISTS source_schedule_id INTEGER REFERENCES task_schedules(id);
+
+-- Add request metadata (IP, user agent) for debugging
+ALTER TABLE worker_tasks
+ADD COLUMN IF NOT EXISTS source_metadata JSONB;
+
+-- Create index for querying by source
+CREATE INDEX IF NOT EXISTS idx_worker_tasks_source ON worker_tasks(source);
+
+-- Comment explaining source values
+COMMENT ON COLUMN worker_tasks.source IS 'Task creation source: schedule, api_run_now, api_crawl_state, api_batch_staggered, api_batch_az_stores, task_chain, manual';
+COMMENT ON COLUMN worker_tasks.source_schedule_id IS 'ID of the schedule that created this task (if source=schedule or source=api_run_now)';
+COMMENT ON COLUMN worker_tasks.source_metadata IS 'Request metadata: {ip, user_agent, endpoint, timestamp}';

backend/migrations/105_dashboard_performance_indexes.sql

@@ -0,0 +1,25 @@
+-- Migration 105: Add indexes for dashboard performance
+-- Purpose: Speed up the /dashboard and /national/summary endpoints
+--
+-- These queries were identified as slow:
+-- 1. COUNT(*) FROM store_product_snapshots WHERE captured_at >= NOW() - INTERVAL '24 hours'
+-- 2. National summary aggregate queries
+
+-- Index for snapshot counts by time (used in dashboard)
+CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_store_product_snapshots_captured_at
+ON store_product_snapshots(captured_at DESC);
+
+-- Index for crawl traces by time and success (used in dashboard)
+CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_crawl_traces_started_success
+ON crawl_orchestration_traces(started_at DESC, success);
+
+-- Partial index for recent failed crawls (faster for dashboard alerts)
+CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_crawl_traces_recent_failures
+ON crawl_orchestration_traces(started_at DESC)
+WHERE success = false;
+
+-- Composite index for store_products aggregations by dispensary
+-- Helps with national summary state metrics query
+CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_store_products_dispensary_brand
+ON store_products(dispensary_id, brand_name_raw)
+WHERE brand_name_raw IS NOT NULL;