Your agent runs on a dedicated VPS that stays powered on 24/7. The dashboard monitors it continuously so you always know its health and status.
Your agent isn't a scheduled job or a serverless function. It runs on its own VPS with continuous monitoring and automatic recovery.
The agent runs on a dedicated Hetzner VPS that stays powered on continuously. Unlike serverless setups that cold-start, your agent is always ready to respond — at 3 AM, on weekends, during holidays.
The dashboard polls your agent every 10 seconds and records a heartbeat with CPU, RAM, disk, and container status. If the agent goes down, you see it immediately on the dashboard.
Heartbeat data is stored in Postgres with 7-day retention, giving you a rolling view of agent health. Telemetry data is retained for 90 days for longer-term analysis.
The VPS API runs as a systemd service with automatic restart on failure. Docker containers are configured with restart policies to recover from crashes without manual intervention.
Every heartbeat records four categories of metrics from your agent's VPS. The data is stored in Postgres and visualized on the dashboard.
Percentage of compute used by the agent and its browser sandbox
Used vs. total memory (bytes), with percentage
Used vs. total disk space, with percentage
Docker service status for gateway and browser (running, stopped, restarting)
Illustrative visualization — actual data from your agent's heartbeats shown on the dashboard
The difference between manually managing tools on your local machine and deploying an always-on agent with built-in monitoring.
A dedicated VPS with automatic health monitoring and crash recovery. Your agent stays available around the clock.