Skip to content

Monitoring & Alerts

Automation only helps if the operator knows what the system is doing and gets warned when something breaks. This page covers the monitoring layer: alerts, dashboards, watchdogs, and health checks that keep Master Rulebook operational.

Telegram Alerts

Telegram is the system's real-time notification channel. These alerts matter because they turn the trading stack from a black box into something you can supervise.

All alerts go to Telegram bot @Ash_clw_bot. Real-time notifications for:

  • Regime changes
  • Plan H/C signal detection
  • Screener results (go/no-go each morning)
  • Plan Alpha early warnings (-3%) and activations (-5%)
  • Health check failures
  • Auth failures and crashes

Daily Dashboard

Think of the dashboard as the daily command center: one place to see the market state, what the screeners found, and whether the infrastructure is healthy.

Generated by daily_digest.py at 1:50 PM PT. Includes:

  • Current regime + SPY 20d return
  • Market breadth
  • Bottom formation candidates
  • Plan M open positions and new signals
  • Plan Alpha status
  • Convergence signals
  • X Expose sentiment summary
  • Health check status

Monitor Watchdog

This is the automatic "is the guard still standing?" check for the live symbol monitors.

Script: monitor_watchdog.sh | Cron: 6:35 AM PT

Verifies SPY and QQQ symbol monitors are running. If a monitor died: 1. Sends Telegram alert 2. Auto-restarts the monitor 3. Logs the restart

Log Watcher

Instead of waiting for a person to notice a failure in a log file, this watcher scans for trouble continuously during market hours.

Script: log_watcher.py | Cron: Every 5 min during market hours

Monitors cron log files for errors, auth failures, and crashes. Sends Telegram alert on detection.

Health Checks

The maintenance job is the system's routine preflight inspection.

Script: maintenance/run_maintenance.sh | Cron: 4:50 AM PT weekdays, 8 AM Sunday

14 automated checks: - Cron jobs running - OVTLYR cookies valid - API credentials working - Disk space - Process health - Log rotation

Dual Venue Execution

Orders are mirrored across two paper/sandbox venues so the system can compare behavior and catch discrepancies.

Venue Timing Purpose
Alpaca Paper Immediate Primary P&L tracking
Tradier Sandbox +15 min delay Secondary validation