Website Health & Performance Monitor: Uptime, Speed, Resources

One call tells you if a site is healthy — and whether it's getting worse. For each URL it measures uptime and timings, audits the page's resources (a waterfall + page weight + by-type breakdown), computes a performance score with concrete fixes, remembers history across runs to catch deterioration and roll up per day/week/month, and gives an overall verdict. Schedule it (e.g. every 5–15 min) and set webhookUrl to be alerted the moment a site goes down, slows, or deteriorates. Keyless core, agent-callable.

🩺 Read-only. It loads the page and the resources the page itself references, exactly as a browser does — no probing or scanning. Recommendations and signals, not a verdict on anyone else's site.

⚡ The full picture, in one report

Group	What you get
Uptime	up, HTTP status, HTTPS, redirects, error when down
Timings	full phase breakdown: DNS, TCP connect, TLS handshake, TTFB, download, total (ms)
TLS cert	days until the certificate expires, and any cert problem
Content	optional keyword assertion - flags content-mismatch when a 200 page is broken/defaced
Resources	request count, total page weight (KB), by-type breakdown (JS/CSS/images), largest + slowest resources, uncompressed/uncacheable counts
Performance	a 0–100 synthetic score + grade + specific issues (slow TTFB, heavy page, uncompressed assets, no HTTPS…)
History	per day/week/month rollups (uptime %, avg load, avg score) and deterioration vs the recent baseline
Verdict	healthy / slow / deteriorating / down
Lighthouse (optional)	real performance score + Core Web Vitals (FCP/LCP/CLS/TBT) via PageSpeed Insights, if you paste your own free key

🔁 Monitoring & alerts

• Schedule the Actor (Apify Schedules) to run every few minutes — it builds history and rollups automatically.
• webhookUrl POSTs any down/slow/deteriorating sites as JSON — wire it to Slack/email/SMS via Make or Zapier.
• Set onlyProblems: true so scheduled runs emit only what needs attention.

📥 Input

{ "urls": ["https://your-site.com"], "auditResources": true, "slowLoadMs": 3000 }

Add pageSpeedApiKey (your own free key) for real Lighthouse + Core Web Vitals. One report per URL.

📤 Example output (real run, trimmed — python.org)

{ "finalUrl": "https://www.python.org/", "signal": "healthy", "status": 200,
  "ttfbMs": 190, "loadTimeMs": 199,
  "performance": { "score": 92, "grade": "A", "issues": ["HTML is not compressed (gzip/br)."] },
  "resources": { "requestCount": 14, "resourceWeightKB": 393,
    "byType": { "script": {"count": 9, "KB": 222}, "stylesheet": {"count": 4, "KB": 155} },
    "largestResources": [ {"type": "stylesheet", "KB": 122} ] },
  "history": { "day": {"checks": 12, "uptimePct": 100.0, "avgLoadMs": 205 } } }

💰 Pricing (pay-per-event)

• Site checked — $0.02 per URL per run (includes the full resource audit). Lighthouse (your key) adds no fee.

💡 Use cases

• Site owners / ops / SRE: uptime + performance + deterioration alerts without standing up a monitoring stack.
• Agencies / MSPs: watch every client site's health and page weight from one place, with history.
• Growth / e-commerce: page weight and load creep quietly hurt conversion — catch the trend early.
• Agents / CI: callable as an MCP tool — give it URLs, get back a health verdict and the numbers.

🔑 Keyless core & agent-callable

The core (uptime, timings, resource waterfall, performance score, history) needs no keys or external services. Real Lighthouse is optional and uses your PageSpeed key.

🚀 Ready-made tasks (one click)

• Website Uptime & Performance Monitor (Scheduled)

🤖 For AI agents (MCP)

https://mcp.apify.com/?tools=actors,docs,centorprat/website-health-performance-monitor

Loads as centorprat--website-health-performance-monitor. Claude Code:

claude mcp add apify -- npx -y @apify/actors-mcp-server --actors centorprat/website-health-performance-monitor