← punyakrit singh makhni

Pulse

site · source

Real-time uptime and performance monitoring with distributed regional checks, false-positive filtering, and automated alerting.

the problem

Most small teams rely on monitoring that is expensive, noisy, or single-region - sending false alerts or missing real downtime because checks only run from one place.

what i built

Designed a distributed monitoring system that runs checks globally, logs real metrics, filters false positives, and sends instant alerts through email, SMS, and webhooks.

why it's interesting

Not an AI project - it is the infrastructure half of the same skill set. Scheduling distributed work, reasoning about partial failure, and deciding when a signal is real are the same problems that show up underneath every agent system I build.

built with

next.js, typescript, postgresql, distributed systems, tailwind css, framer motion