What is Watchdog?
Watchdog is an event-driven uptime monitoring service that periodically checks configured URLs, writes time-series metrics to TimescaleDB/Postgres, and notifies owners when status changes occur. Built with Go, it provides a robust and scalable solution for monitoring your critical web services. The system uses a sophisticated worker-based architecture where child workers perform HTTP checks, a supervisor evaluates results and makes state decisions, and event listeners handle persistence and notifications—ensuring clean separation of concerns throughout the monitoring pipeline.Key features
Event-driven architecture
Built on a pub/sub event bus that decouples check execution from persistence and notifications, enabling flexible and maintainable monitoring workflows.
Flexible monitoring intervals
Monitor URLs at intervals from 10 seconds to 24 hours. Choose the frequency that matches your service’s criticality and requirements.
Time-series metrics
Stores historical health and latency data in TimescaleDB hypertables, giving you powerful analytics and insights into your services over time.
Instant notifications
Receive email alerts via SMTP when your monitored URLs change state, ensuring you’re always aware of issues as they happen.
Multiple HTTP methods
Monitor endpoints using GET, POST, PATCH, PUT, or DELETE methods to accurately test your API health.
Simple CLI interface
Manage your monitored URLs with intuitive commands: add, remove, list, and analyze—all from your terminal.
How it works
Watchdog’s architecture is designed for reliability and scalability:Orchestrator initialization
The orchestrator bootstraps the system by setting up the logger, event bus, and supervisor. It then registers event listeners and creates parent worker groups for each monitoring interval.
Worker execution
Parent workers spawn child workers that perform the actual HTTP checks. Each child worker makes requests to assigned URLs based on the configured interval and forwards raw results to the supervisor.
Supervisor evaluation
The supervisor receives check results and applies decision logic (thresholds, debounce) to determine if a check represents success or failure, then publishes domain events like
ping.successful or ping.unsuccessful to the event bus.The event-driven design ensures that workers focus solely on performing checks, while the supervisor handles state management, and listeners manage side effects. This separation of concerns makes the system maintainable and extensible.
Tech stack
Watchdog is built with modern, production-ready technologies:- Go 1.24 - Fast, concurrent, and reliable
- PostgreSQL with TimescaleDB - Time-series data storage via
pgx/v5 - Redis - Internal caching and work distribution
- Goose - Database migration management
- SMTP - Email notifications (Mailtrap, SendGrid, etc.)
- Structured logging - Built-in
log/slogfor observability
Get started
Quick start
Get Watchdog up and running in minutes with our step-by-step guide
CLI reference
Explore all available commands and their options