PulseWatch — Amresh Chaurasiya

Overview

PulseWatch is an observability platform built around a robust distributed worker pipeline. It schedules continuous URL probes, manages complex incident state transitions with hysteresis, and dispatches multi-channel alerts. The architecture guarantees high-availability polling without relying on external SaaS providers.

Architecture

01 — Dashboard

Next.js UI for status pages and metrics

02 — Scheduler

Cron-based heartbeat trigger

03 — Redis Streams

Distributed, reliable job queue

04 — Monitoring Workers

Global HTTP/TCP probe executors

05 — PostgreSQL

Historical latency and incident logs

06 — Incident Engine

State machine for failure handling

07 — Alert Engine

Webhook and email dispatch

Why I built this

Reliability monitoring shouldn't require sending internal API layouts to third-party services. I wanted to build an incident management system from first principles to deeply understand distributed job processing, state machines, and reliable alerting.

Engineering Challenges

Building a distributed job pipeline using Redis Streams that guarantees at-least-once delivery
Incident state machine hysteresis: distinguishing between a temporary network blip and a hard failure to prevent alert fatigue
Handling monitoring worker node failures seamlessly

What I learned

Reliability in a monitoring system is recursive — the tool that watches your uptime has to be more reliable than the things it watches.

Redis Streams consumer groups provide an incredibly robust primitive for reliable worker pipelines compared to standard Pub/Sub.

Stack

Next.js

TypeScript

Node.js

Redis Streams

PostgreSQL

Prisma

DockerResend