WatchTower
Built a self-hosted, asynchronous uptime and API performance monitoring engine leveraging Celery workers, Redis brokers, FastAPI, and Next.js for high-frequency telemetry.
99.99%
Uptime Tracking
Granular second-by-second availability
Async Celery
Worker Concurrency
Non-blocking event loop execution
Up to 10s
Metrics Frequency
Configurable ping intervals per endpoint
100%
Self-Hosted
Zero reliance on expensive third-party SaaS
Operational Constraints & Friction
Monolithic commercial performance suites are cost-prohibitive, opaque, and overly complex for decentralized endpoint health, latency tracking, and API uptime diagnostics.
Architectural Execution
Built a self-hosted, asynchronous uptime and API performance monitoring engine leveraging Celery workers, Redis brokers, FastAPI, and Next.js for high-frequency telemetry.
WatchTower — System Architecture
High-Level Topology & Data Flow
Next.js 15 Real-time Dashboard & Telemetry Visualizer
Client requests, sensors & telemetry
FastAPI REST Core & WebSocket Bridge
ACID Persistence & Read/Write separation
Internal Topology & Data Mechanics
Celery Beat schedules asynchronous HTTP ping tasks through a Redis broker to Celery worker pools. Results are aggregated, buffered, and committed to PostgreSQL, while a Next.js frontend fetches telemetry via FastAPI REST endpoints.
Engineering Obstacles Overcome
Managing concurrent background network workers smoothly without blocking main application routing loops or overwhelming the relational database with high-frequency check logs.
- ✓Asynchronous distributed uptime checks
- ✓Interactive real-time telemetry dashboard
- ✓Histogram latency metrics & percentile reporting
- ✓Multi-target endpoint health scoring
- ✓Configurable ping frequencies & failure alerting
Planned Enhancements & Roadmap
Project Delivery & Impact
Engineered a lightweight, non-blocking monitoring microservice suite that safely schedules high-frequency telemetry requests across distributed targets.