docs: update API docs for async download behavior

Document:
- New downloading_data status
- Warnings field in responses
- Async flow and monitoring
- Example usage patterns

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

API_REFERENCE.md

@@ -269,12 +269,14 @@ Get status and progress of a simulation job.
 | `total_duration_seconds` | float | Total execution time in seconds |
 | `error` | string | Error message if job failed |
 | `details` | array[object] | Per model-day execution details |
+| `warnings` | array[string] | Optional array of non-fatal warning messages |
 
 **Job Status Values:**
 
 | Status | Description |
 |--------|-------------|
 | `pending` | Job created, waiting to start |
+| `downloading_data` | Preparing price data (downloading if needed) |
 | `running` | Job currently executing |
 | `completed` | All model-days completed successfully |
 | `partial` | Some model-days completed, some failed |
@@ -289,6 +291,35 @@ Get status and progress of a simulation job.
 | `completed` | Finished successfully |
 | `failed` | Execution failed (see `error` field) |
 
+**Warnings Field:**
+
+The optional `warnings` array contains non-fatal warning messages about the job execution:
+
+- **Rate limit warnings**: Price data download hit API rate limits
+- **Skipped dates**: Some dates couldn't be processed due to incomplete price data
+- **Other issues**: Non-fatal problems that don't prevent job completion
+
+**Example response with warnings:**
+
+```json
+{
+  "job_id": "019a426b-1234-5678-90ab-cdef12345678",
+  "status": "completed",
+  "progress": {
+    "total_model_days": 10,
+    "completed": 8,
+    "failed": 0,
+    "pending": 0
+  },
+  "warnings": [
+    "Rate limit reached - downloaded 12/15 symbols",
+    "Skipped 2 dates due to incomplete price data: ['2025-10-02', '2025-10-05']"
+  ]
+}
+```
+
+If no warnings occurred, the field will be `null` or omitted.
+
 **Error Response:**
 
 **404 Not Found** - Job doesn't exist

docs/user-guide/using-the-api.md

@@ -94,6 +94,70 @@ curl "http://localhost:8080/results?job_id=$JOB_ID&date=2025-01-16&model=gpt-4"
 
 ---
 
+## Async Data Download
+
+The `/simulate/trigger` endpoint responds immediately (<1 second), even when price data needs to be downloaded.
+
+### Flow
+
+1. **POST /simulate/trigger** - Returns `job_id` immediately
+2. **Background worker** - Downloads missing data automatically
+3. **Poll /simulate/status** - Track progress through status transitions
+
+### Status Progression
+
+```
+pending → downloading_data → running → completed
+```
+
+### Monitoring Progress
+
+Use `docker logs -f` to monitor download progress in real-time:
+
+```bash
+docker logs -f ai-trader-server
+
+# Example output:
+# Job 019a426b: Checking price data availability...
+# Job 019a426b: Missing data for 15 symbols
+# Job 019a426b: Starting prioritized download...
+# Job 019a426b: Download complete - 12/15 symbols succeeded
+# Job 019a426b: Rate limit reached - proceeding with available dates
+# Job 019a426b: Starting execution - 8 dates, 1 models
+```
+
+### Handling Warnings
+
+Check the `warnings` field in status response:
+
+```python
+import requests
+import time
+
+# Trigger simulation
+response = requests.post("http://localhost:8080/simulate/trigger", json={
+    "start_date": "2025-10-01",
+    "end_date": "2025-10-10",
+    "models": ["gpt-5"]
+})
+
+job_id = response.json()["job_id"]
+
+# Poll until complete
+while True:
+    status = requests.get(f"http://localhost:8080/simulate/status/{job_id}").json()
+
+    if status["status"] in ["completed", "partial", "failed"]:
+        # Check for warnings
+        if status.get("warnings"):
+            print("Warnings:", status["warnings"])
+        break
+
+    time.sleep(2)
+```
+
+---
+
 ## Best Practices
 
 ### 1. Check Health Before Triggering