Topic hub

Website Monitoring

Website monitoring should detect failures that affect customers or business operations, not merely prove that the home page answered one automated request. Combine availability, errors, performance and transaction…

1 published guides0 subtopicsUpdated 21 July 2026

Start here

The main guide for this topic

Practical guide

Broken Link Checking

Broken links waste users’ time, hide maintenance problems and can prevent important pages or downloads from being discovered. Automated reports are useful, but every result needs context before a redi…

Open the complete guide →

Published guides

Articles in this section

Practical guide22 September 2026

Broken Link Checking

Broken links waste users’ time, hide maintenance problems and can prevent important pages or downloads from being discovered. Auto…

Read guide →

About this section

Information checked: 21 July 2026.

In brief: Monitor the customer journey and make every alert actionable.

Monitor in layers

Monitoring layers
Layer	Example signal	Business meaning
Availability	DNS, TLS and HTTP response	Can a visitor reach the service?
Application	Error rate, queue failure and resource exhaustion	Is the system functioning internally?
Transaction	Form, booking, login or checkout test	Can a customer complete the task?
Performance	Field and synthetic timings	Is the task usable under real conditions?
Integrity and security	Unexpected changes or privileged access	Has the service changed without authorisation?

Design useful alerts

State exactly what failed and from where.
Include the first observed time and recent changes.
Route critical alerts to more than one verified channel.
Group repeated symptoms into one incident.
Define who acknowledges, investigates and escalates.
Test alert delivery and out-of-hours handling.

Control noise

An alert that fires constantly without action becomes invisible. Tune thresholds using business impact, suppress expected maintenance and review alerts that were ignored or failed to detect a real incident.

Acceptance check

Monitoring is ready when a controlled failure creates an understandable alert, reaches the responsible person and leads to a documented response.

Monitor from the customer’s point of view

A server dashboard can remain green while form notifications are rejected or a booking provider fails. Add at least one end-to-end check that confirms the customer action and the business-side receipt. Use test records that staff can recognise and remove safely.

Records to keep

Monitor inventory and business priority.
Alert routing and escalation matrix.
Synthetic-test accounts and cleanup procedure.
Monthly alert-quality review.

Owner test: Disable one non-production dependency and verify that the alert identifies the failed customer journey rather than only reporting a generic timeout.

Review the monitor itself

Monitoring accounts, API keys and notification channels can expire or be removed during staff changes. Schedule a periodic self-test that confirms checks are still running, alerts reach current responders and the monitoring provider is not relying on the same infrastructure it watches.

Sources and date checked

This practical guidance was checked against the following primary sources. Date checked: 21 July 2026.