Your website could be down right now and you wouldn't know — unless you have monitoring set up. Studies show the average website experiences 3-5 hours of downtime per month, and every minute of downtime costs businesses money, credibility, and search rankings.
Website monitoring is one of those things that takes 10 minutes to set up but saves you hours of damage control. Here's how to do it right.
Why Monitoring Matters
Without monitoring, you rely on customers to tell you your site is down. By then, you've already lost sales, frustrated visitors, and potentially dropped in Google rankings.
- Revenue loss: E-commerce sites lose an average of $5,600 per minute of downtime
- SEO impact: Google notices downtime and may temporarily lower your rankings
- Trust erosion: Visitors who hit a dead site may never come back
- SLA compliance: If you promise 99.9% uptime to clients, you need proof
Types of Website Monitoring
1. Uptime Monitoring
The most basic and essential type. A service pings your website at regular intervals and alerts you if it doesn't respond.
What it checks:
- Is the server responding?
- Is it returning the correct HTTP status code (200 OK)?
- Is it responding within an acceptable time?
Free tools:
UptimeRobot (recommended for getting started):
- 50 monitors free
- 5-minute check intervals
- Email, SMS, Slack, and webhook alerts
- Status pages included
Setup takes 2 minutes:
- Sign up at uptimerobot.com
- Click "Add New Monitor"
- Select "HTTP(s)" monitor type
- Enter your URL
- Set alert contacts
- Done
Other free options:
- Freshping — 50 monitors, 1-minute intervals
- Hetrix Tools — 15 monitors, 1-minute intervals
- StatusCake — 10 monitors, 5-minute intervals
2. Performance Monitoring
Goes beyond "is it up?" to answer "is it fast?" Performance monitoring tracks page load times, server response times, and Core Web Vitals.
Key metrics to track:
- TTFB (Time to First Byte): How fast your server responds. Target: under 200ms
- FCP (First Contentful Paint): When the first content appears. Target: under 1.8s
- LCP (Largest Contentful Paint): When the main content loads. Target: under 2.5s
- CLS (Cumulative Layout Shift): Visual stability. Target: under 0.1
Tools:
Google PageSpeed Insights — Free, on-demand testing at pagespeed.web.dev
Google Search Console — Free, shows Core Web Vitals for your actual visitors over time. Essential for SEO.
GTmetrix — Free tier with detailed waterfall analysis showing exactly what slows your page down.
3. Real User Monitoring (RUM)
Tracks performance as experienced by actual visitors, accounting for different devices, browsers, and locations.
Google Analytics provides basic RUM data for free:
- Go to Behavior → Site Speed → Overview
- See average page load times by page, browser, and country
For more detailed RUM, consider Cloudflare Web Analytics (free) or New Relic Browser (free tier available).
4. Server Monitoring
Tracks what's happening on your server — CPU, RAM, disk usage, and network traffic. Critical for VPS and dedicated server users.
# Quick server health check script
#!/bin/bash
echo "=== Server Health ==="
echo "CPU Load: $(uptime | awk -F'load average:' '{print $2}')"
echo "Memory: $(free -h | awk '/^Mem:/{print $3 "/" $2}')"
echo "Disk: $(df -h / | awk 'NR==2{print $3 "/" $2 " (" $5 " used)"}')"
echo "Uptime: $(uptime -p)"
Tools:
- Netdata — Free, real-time server monitoring with a beautiful dashboard. Install in one command:
bash <(curl -Ss https://my-netdata.io/kickstart.sh)
- htop — Terminal-based process viewer (already on most servers)
- Prometheus + Grafana — Industry standard for advanced monitoring (free, self-hosted)
5. SSL Certificate Monitoring
SSL certificates expire. When they do, visitors see scary "Your connection is not private" warnings. Monitor certificate expiry to avoid this.
# Check SSL expiry from command line
echo | openssl s_client -servername yourdomain.com -connect yourdomain.com:443 2>/dev/null | openssl x509 -noout -dates
Most uptime monitoring tools (UptimeRobot, StatusCake) include SSL expiry alerts automatically.
Setting Up Alerts That Actually Work
Monitoring is useless without good alerts. Here's how to avoid alert fatigue while catching real issues:
Alert Channels
Set up multiple notification channels for redundancy:
- Email — For non-urgent alerts and daily summaries
- SMS/Phone — For critical downtime (your site is DOWN)
- Slack/Discord — For team visibility
- PagerDuty/Opsgenie — For on-call rotation (larger teams)
Alert Strategy
- Confirm before alerting: Most tools wait for 2-3 consecutive failures before alerting. This prevents false alarms from network blips.
- Escalation: If the site is still down after 15 minutes, escalate to a phone call.
- Downtime threshold: Alert immediately for complete downtime. For slow performance, alert only if it persists for 5+ minutes.
- Maintenance windows: Suppress alerts during planned maintenance to avoid unnecessary noise.
Creating a Public Status Page
A status page shows your users the current state of your services. When something goes wrong, visitors check the status page instead of flooding your support inbox.
Free options:
- UptimeRobot Status Pages — Included with free plan
- Instatus — Free tier with custom domain
- Cachet — Self-hosted, open source
A good status page shows:
- Current status of each service (website, API, email, etc.)
- Uptime percentage over 30/90 days
- Incident history with updates
- Planned maintenance schedule
Monitoring Checklist
Here's what every website should monitor at minimum:
- Homepage uptime — HTTP check every 5 minutes
- Key pages — Monitor login, checkout, API endpoints separately
- SSL certificate — Expiry alerts at 30, 14, and 7 days
- Domain expiry — Don't let your domain lapse
- Server resources — CPU, RAM, disk alerts at 80% threshold
- Response time — Alert if TTFB exceeds 1 second consistently
- Error rates — Monitor 5xx errors in server logs
Automated Response Scripts
For common issues, you can set up automated recovery:
#!/bin/bash
# auto-restart.sh — Restart app if health check fails
HEALTH_URL="http://localhost:3000/health"
RESPONSE=$(curl -s -o /dev/null -w "%{http_code}" $HEALTH_URL)
if [ "$RESPONSE" != "200" ]; then
echo "$(date): Health check failed (HTTP $RESPONSE). Restarting..."
pm2 restart my-app
# Or: systemctl restart nginx
# Or: docker compose restart app
fi
Run it via cron every minute:
* * * * * /path/to/auto-restart.sh >> /var/log/auto-restart.log 2>&1
Interpreting Your Data
Once monitoring is running, review your data weekly:
- Uptime trend: Are you meeting 99.9% (8.7 hours downtime/year)?
- Performance trend: Are load times increasing? Could indicate growing traffic or bloated code.
- Error patterns: Do errors spike at certain times? Could be cron jobs, traffic peaks, or resource limits.
- Geographic performance: Is your site slow in certain regions? Consider a CDN.
Start Monitoring Today
Website monitoring is non-negotiable for any serious online presence. The good news is that getting started takes minutes, not hours, and many excellent tools are completely free.
At DeployBase, our hosting plans include built-in server monitoring with CPU, memory, and disk alerts out of the box. Combined with automatic daily backups and 99.9% uptime SLA, you get peace of mind that your site is always running at its best. Focus on building your business — we'll watch the servers.
