📈 Monitoring Operations
Algorithmic trading demands high-integrity, real-time observability. Because Hoox microservices are distributed across Cloudflare’s global edge network, tracking health, logs, and message queues requires a consolidated management plane.
This guide outlines how to use the Hoox CLI to audit system status, monitor queue backlogs, stream live edge logs, and manage the emergency kill switch.
🟢 1. Probing Worker Health Status
Every Hoox worker is configured with a secure /health endpoint that validates local CPU capacity, binding access, and D1 database connection integrity.
To probe and audit all active endpoints simultaneously:
hoox monitor status was removed in v0.8.0. The single source of truth for health checks is now hoox check health.
Expected Output
The CLI runs asynchronous concurrent health probes, displaying green indicators and public route locations:
Interpreting Health States
🚨 2. Governing the Global Kill Switch
The Global Kill Switch is your emergency brake. Stored inside the sub-millisecond CONFIG_KV namespace, flipping this parameter instantly blocks all incoming trade signals globally in under 10 seconds, without requiring code updates or worker redeployments.
Kill Switch Behavior
When activated, the following sequence occurs automatically:
hoox gateway rejects all incoming /webhook POST
requests with 503 Service Unavailable.- A
KILL_SWITCH_ACTIVE event is logged to D1 for audit trail. agent-worker halts its 5-minute cron execution loop.trade-worker queues any in-flight payloads for retry (not
abandoned).- Telegram notification is sent to alert you.
When the Kill Switch is turned ON, the hoox gateway router immediately
rejects all incoming TradingView alerts and API webhook requests with a 503 Service Unavailable error and logs a KILL_SWITCH_ACTIVE warning. Active
positions must be flattened manually or via hoox monitor trades close-all.
🗃️ 3. Auditing Recent Transactions
To review execution history directly from your terminal:
This aggregates entries from your production D1 database and prints a formatted terminal table outlining: TIMESTAMP | REQUEST_ID | EXCHANGE | SYMBOL | ACTION | QUANTITY | PRICE | STATUS
Trade Status Codes
📨 4. Inspecting Queue Depth (Guaranteed Delivery Backlog)
If network volatility causes exchange API dropouts, Hoox offloads trade execution payloads to Cloudflare Queues. To check if there is an active execution backlog:
This displays the number of pending messages in the queue:
0: System is running normally (all trades executing on the fast-path direct service binding).> 0: Exchange API is experiencing rate-limits or downtime. Trade-worker is retrying transactions asynchronously in the background.
Queue Health Indicators
📝 5. Real-Time Log Streaming & Telemetry
When debugging custom strategies, you can stream verbose console logs directly from Cloudflare’s global edge nodes to your local terminal using Wrangler bindings:
Downloading Logs for Off-Line Analysis
To download historical logs offloaded to your R2 storage bucket for compliance audits or tax reports:
📊 6. System Metrics Overview
Hoox tracks key performance indicators via Cloudflare Analytics Engine. Access summary metrics via:
Key Metrics Tracked
If the system status check returns a red error indicator (e.g. d1-worker ... ❌ FAILED), immediately proceed to the Self-Healing & Repair
Guide to run automated diagnostics and restore service bindings! 🔗 Next Steps
