The eG Approach to Real User Monitoring
eG Enterprise employs a proven, time-tested, ‘Java script’ approach to monitoring real user transactions to web sites/web applications. This is an ‘agentless approach’ that involves embedding a small Java script in every page to be monitored in the target web site/web application. Whenever the browser loads one of these web pages in response to a user request, the script runs and collects the following metrics:
- The page load time;
- Break-up of page load time;
- Count of Java script errors (if any);
- Which browser was used to access the page? What is the browser version?
- On what operating system is the browser running?
- From which device the request came in – desktop? mobile phone? or tablet?
Upon metrics collection, the browser sends performance beacons carrying the gathered metrics to a software component called the eG RUM Collector.
Figure 1 : How the eG RUM works
The collector receives the metrics and stores them locally in flat files. Additionally, the collector also intelligently determines the geography (country, city, and region) from where the request originated and stores this data too in the flat files. An eG agent deployed on a remote host, can then be configured to poll the RUM collector at regular intervals to collect the request and response time metrics and the details of the geography. The remote eG agent then aggregates the collected metrics based on:
- The web site/web application monitored
- URL groups configured for monitoring
- Transaction URL patterns configured for monitoring
- Geographies – i.e., countries, cities, and regions
- Page types – Page, Iframe, AJAX
The aggregated results are then compared with thresholds to isolate deviations. Both the aggregated metrics and the threshold violations are then reported to the eG manager, which presents the performance and problem information in the eG monitoring console and also stores the same in the eG backend.