←  All terms
M
QA & Deployment

Monitoring

Monitoring continuously observes systems and user experiences through metrics, logs, and alerts.

Monitoring continuously observes systems and user experiences through metrics, logs, and alerts. Automated checks speed up validation but do not replace focused manual testing of critical user journeys. The concept is closely related to Uptime and Real User Monitoring.

Read more

Uptime

Real User Monitoring

Resilience

Service Level Agreement

FAQ

Monitoring

How does monitoring relate to a service level agreement?
An SLA sets target values such as availability, and monitoring supplies the measurements against which compliance can be checked. Without a defined measurement method, any commitment stays vague. Which figures apply contractually and what follows from deviations is a matter of the contract, not of the monitoring tool.
When is simple uptime monitoring no longer enough?
Uptime monitoring only shows whether a page responds, not whether it works. Once forms, logins, integrations, or performance are business-critical, synthetic tests of individual user journeys and real user monitoring are needed on top. A reachable page with a broken contact form still counts as an outage for visitors.
How can false alarms in monitoring be avoided?
Alerts should be tied to user impact, not to isolated fluctuations. Repeated checks from several locations, sensible thresholds, and clear ownership reduce noise. Teams that receive too many notifications eventually ignore the relevant ones as well. The rule of thumb: fewer alerts, each with a clear action attached.
What should be monitored on a website?
At a minimum: availability, response times, error rates, and critical user journeys such as forms or checkout. Certificate expiry, domain validity, and field data on Core Web Vitals come on top. The scope follows business risk: whatever costs revenue or reputation when it fails deserves the tightest coverage.
What is the difference between monitoring and testing?
Testing checks before release whether a function works as expected. Monitoring observes in live operation whether it keeps doing so. A passed test says nothing about later outages, expiring certificates, or behavior under real load. The two disciplines cover different points in time and complement each other.