Website performance and health monitoring: what to measure and when to alert
Response time, percentiles, TTFB and Core Web Vitals: how website performance monitoring works, synthetic vs real-user data, and the health checks that catch problems weeks early.
· 4 min read · By the Spot Downtime team
A site that takes eight seconds to load is up as far as an uptime check is concerned, and losing visitors all the same. Website performance monitoring tracks how fast your pages are over time, so a slowdown gets noticed when it starts, not when customers complain. Website health monitoring adds the slow-moving risks around it: certificates, domains and DNS.
Synthetic vs real-user monitoring
| Synthetic | Real-user (RUM) | |
|---|---|---|
| How | A robot loads the page on a schedule | A script on your pages reports what visitors experience |
| Strengths | Consistent, works with no traffic, catches problems at 3 a.m. | Real devices, networks and countries |
| Blind spots | One device and network; not your real audience | Needs traffic; says nothing when nobody visits |
| Best for | Alerting and trends | Understanding and prioritising fixes |
Uptime monitors are synthetic. They're the right tool for alerting because they run whether or not anyone is visiting. RUM is the right tool for deciding what to optimise.
What to measure
Response time, as percentiles
Averages hide pain. If 19 requests take 100 ms and one takes 3 seconds, the average is 245 ms, a number that describes none of them, while one visitor in twenty waits 3 seconds. Watch the p95 (the time 95% of requests beat) and p99 alongside the median. Spot Downtime records every check's response time and shows p50, p75, p95 and p99 for each monitor.
Time to first byte
TTFB is how long the server takes to start answering. It's the part you control on the back end: slow database queries, cold caches, an overloaded server. A rising TTFB with an unchanged page is almost always a server-side problem.
Core Web Vitals
Google's three page-experience metrics, measured in a real browser:
- LCP (Largest Contentful Paint): when the main content appears. Good is 2.5 seconds or less.
- INP (Interaction to Next Paint): how quickly the page responds to taps and clicks. Good is 200 ms or less.
- CLS (Cumulative Layout Shift): how much the page jumps around while loading. Good is 0.1 or less.
LCP Main content shown
1.9 s · GoodINP Response to a tap
340 ms · Needs improvementCLS Layout shift
0.04 · GoodThese need a full browser load, so they're measured less often than uptime: on each release, or weekly. The website speed test runs Lighthouse and shows all three with a filmstrip and request waterfall.
Alert on slowness, not just downtime
The useful alert is “slower than normal for a while”, not “one slow request”. On Starter and Pro, a Spot Downtime slowness alert fires when a monitor's p95 response time stays above a threshold you set, without counting it as downtime. Set the threshold a little above your normal p95, so it fires on a real change rather than every busy hour. The response time guide shows how to pick it.
Website health monitoring
Performance is one part of a site's health. The rest are checks that catch problems days or weeks before they become outages:
| Check | Catches | In Spot Downtime |
|---|---|---|
| SSL certificate expiry | A renewal that silently failed | Warnings 14, 7 and 1 day before |
| Domain expiry | A lapsed card on the registrar account | Warnings 30, 7 and 1 day before |
| DNS records | A record changed or deleted by mistake | DNS monitors |
| Keyword on the page | A 200 that serves an error page | Page must contain |
| Background jobs | Backups and queues that stopped | Heartbeats |
| Broken links | Pages and assets that 404 | Broken link checker |
What we don't do
Get new guides by email
Practical posts on monitoring, outages, error fixes and on-call, when they're published. No spam.
Keep reading
- Monitoring · Servers · UptimeKubernetes liveness and readiness probes vs external uptime monitoringProbes keep pods healthy inside the cluster. They can't tell you users can't reach the site. How liveness, readiness and startup probes work, and what to add outside.November 15, 2026 · 7 min read
- Monitoring · Uptime · AlertsHow often should an uptime monitor check your site?Check interval decides how long an outage goes unnoticed. How to pick between 30 seconds, 1 minute and 5 minutes, and which services deserve the faster setting.November 14, 2026 · 6 min read