> Proactive Monitoring
How Morpheus watches your system and alerts you to problems
Overview
Morpheus includes a proactive monitoring daemon that runs in the background on your machine. Instead of waiting for you to ask “how's my disk space?”, it continuously watches key metrics and pushes alerts to your phone when something needs attention.
Built-in Watchers
System Health Monitor
Tracks CPU usage, memory usage, and disk space. Configurable thresholds (default: CPU 90%, memory 85%, disk 90%). Alerts include suggested actions and have a 5-minute cooldown to prevent notification spam.
File System Monitor
Watches configured directories for changes. Useful for monitoring config files, deployment directories, or shared folders. Supports include/exclude patterns and debouncing.
Process Monitor
Tracks named processes and alerts when they stop or start. Perfect for monitoring database servers, web servers, or any critical daemon.
MCP Health Monitor
Watches your installed MCP servers. If one crashes, Morpheus auto-restarts it (up to 3 attempts) and notifies you.
Self-Healing
Morpheus doesn't just detect problems — it fixes them. The Health Manager supervises all agent services (WebSocket, Claude CLI, tunnel, MCP servers) and auto-restarts any that crash. It uses exponential backoff and tracks restart counts to avoid infinite loops.
The Failure Tracker records every error with classification (rate limit, timeout, network, tool error, etc.) and builds adaptive retry strategies. Instead of fixed 1s/2s/4s delays, Morpheus learns the optimal retry timing for each error type from historical data.
Configuration
All watchers are disabled by default. Enable them from the mobile app's settings:
- Open Morpheus on your phone
- Go to Settings → Monitoring
- Enable the watchers you want
- Configure thresholds to your preferences
Watcher configuration is persisted on the agent and survives restarts. Notifications are queued when your phone is disconnected and delivered on reconnect.