Monitoring

Rapid identification of problems.

Drupal monitoring overview

Aim

Ensure that the production site is up and running and operating correctly. Avoid your users being treated as unpaid testers of your software.

Approach

Monitor the website with external tools which are able to either check in and report on the health or the site or receive messages when errors occur.

Outcome

When things go wrong on a monitored site the problem is logged and a correcting step is executed or a human is notified. Without this monitoring the problem will most likely go unnoticed... until users start complaining.

Drupal website monitoring

At Morpht we monitor websites in a number of ways to ensure that we keep abreast of their health and are notified as soon as there is an issue.

Uptime monitoring: The highest level of monitoring which checks to see if the site is up and running. Monitoring can be simple and limited to a page returning a 200 response. A page can also be checked for returning a certain result to ensure that it is operational.

Error logging: The application or webserver will emit error messages from time to time. It is vital that these are logged into a system where they will be retained for later analysis. We use tools such as Kibana to review logs of this nature, or Logstash to detect errors in logs and alert or take a corrective action automatically.

Search Console: Google offers a search console which will report on content problems on the site. This is helpful for identifying broken links and content issues.

Site quality tools: There are a number of providers of SaaS tools which are able to monitor the site for quality. Such tools include Monsido and SiteImprove. Both of these tools can be helpful for large organisations where monitoring of quality of content across a number of pages is vital.

Frequently asked questions

Uptime and availability, performance (page load times, error rates), security (intrusion detection, vulnerability alerts), and increasingly accessibility/quality regressions caught by automated tools; comprehensive monitoring means knowing about a problem before users report it, not after.

A site assessment is a point-in-time diagnostic; monitoring is continuous and automated, catching issues as they emerge in real time. Both matter; assessment for periodic deep review, monitoring for day-to-day operational assurance, and they're complementary rather than substitutes for one another.

This depends entirely on the support arrangement in place; the value of monitoring is limited unless there's a clear, agreed incident response process (who's alerted, response time commitments, escalation path) behind it, which should be an explicit part of any support and maintenance agreement.

Insights

Cypress and CI
Article

This guide explains how to automate your Cypress tests for a Drupal site using GitHub Actions. We’ll cover setting up a full CI environment, connecting a MySQL database, running the Drupal server, and executing end-to-end Cypress tests - all inside GitHub’s cloud runners. You'll learn practical tips to speed up builds with caching and securely handle environment configurations.
Beer poured from clean pipelines
Article

The Morpht CI pipeline caught a recent vulnerability in Drupal core which led to the problem promptly being fixed.
A colourful image representing analytical data visualisation
Article

Google Analytics has recently introduced a new update. As a result of this update, Universal Analytics (the current system) will be retired on 1 July 2023.