Pulse
Real-Time API Uptime & Incident Monitoring
Screenshots
Overview
A production uptime monitor built around a multi-worker job scheduler running on MongoDB primitives rather than a queue library — any number of instances share one queue with no double-execution, and incidents stream to the dashboard as they open and resolve.
Technical Details
Built a multi-worker job scheduler on MongoDB using an atomic findOneAndUpdate as a distributed lock, with a worker_threads pool for off-event-loop HTTP checks — any number of instances can share one queue with zero double-execution.
Incidents auto-open and resolve with MTTR streamed to the React dashboard over WebSockets; uptime and latency analytics are computed in MongoDB aggregation pipelines rather than in application code.
Shipped with a full Jest suite, Docker, and GitHub Actions CI, deployed and running in production.