MONITORING.md

📈 Monitoring Guide

Set up metrics, logs, and alerts to track your trading system

🛠️ Prerequisites

• Prometheus server
• Grafana
• Access to application logs in logs/

🔍 Metrics Export

1. The application exposes Prometheus metrics on port 8000 by default.
2. Configure Prometheus to scrape:

   scrape_configs:
     - job_name: 'trading_system'
       static_configs:
         - targets: ['localhost:8000']

📊 Grafana Dashboards

1. Import our sample dashboard JSON into Grafana (file: monitoring/grafana_dashboard.json).
2. Key panels:
   • Tick Latency: p50/p99 latency of tick processing
   • Signal Rate: Signals generated per minute
   • Order Errors: Failed order count

📝 Log Aggregation

• Logs are written to logs/trading.log in JSON format.
• To centralize logs, forward to a log collector (e.g., Loki, ELK):

  # Example using promtail (Loki client)
  promtail --config.file=promtail-config.yaml

🚨 Alerting

1. Create Prometheus alerts:

   alert: HighTickLatency
   expr: tick_latency_p99 > 0.02
   for: 1m
   labels:
     severity: warning
   annotations:
     summary: "High tick processing latency"
     description: "p99 > 20ms"

2. Configure Alertmanager for email or Slack notifications.

---

Monitoring gives you confidence and early warnings if something goes wrong.