SiteSentinel: Uptime & Web Vitals Monitor

The Problem

Traditional uptime pingers only test if a web server returns an HTTP 200 status code. They completely miss client-side JavaScript crashes, third-party script blocking, degraded Core Web Vitals that hurt Google search rankings, or sudden server response (TTFB) latency spikes. By the time a client reports an issue, days of conversions and SEO visibility have been lost.

I architected SiteSentinel to run true synthetic end-to-end telemetry: spin up headless Playwright browser workers in containerized cloud micro-instances, simulate real user sessions, compute real Web Vitals metrics, and alert engineering teams via webhook within 30 seconds of degradation.

Architecture & Observability Pipeline

Built for reliability and low operational cost:

1. Scheduled Playwright Workers: Dockerized Node.js cron workers dispatch headless Chrome instances across multiple cloud regions (East US, Europe, Asia).
2. Core Web Vitals Telemetry: Intercepts Chrome DevTools Protocol performance timings to measure Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS).
3. SSL & DNS Sentinel: Checks SSL certificate expiration countdown and DNS propagation anomalies.
4. Incident Escalation Router: Dispatches multi-channel alerts (Slack, Discord, SMS via Twilio) when error budgets are breached.

The Stack

Key Capabilities

• Multi-Region Synthetic Probing: Validates global latency from North America, Europe, and Asia.
• Real-Time CWV Scoring: Directly correlates Google Core Web Vitals to SEO ranking impacts.
• Automated Root-Cause Snapshots: Captures full DOM console error logs and network request waterfall charts on failure.