or you are already there, but facing some of these issues?: impact on the user experience of the monitored application, significant network traffic, lack of precision/insights on the actual user issue….??
1. Performance API to collect resource URLs… (rather than scanning and tracking changes to the DOM for them)
2. DOM only when it makes sense
3. Assess/Reduce the impact (response, memory and network traffic). Here is an idea of how much data is stored within germainAPM database while monitoring an eCommerce site that has 1000 concurrent users and depending on how heavily these users browse the eCommerce site:
4. Asynchronous Transaction as much as possible
5. Away from the Main Thread whenever possible
6. And focus on providing metrics that will help your developers understand and fix issues, such as:
Breakdown the user click response to understand how much time is spent on the browser, network, infrastructure and application.
i.e. collect in real-time: User Click, Move, Page response time are broken down and includes the following:
Network Time/Server Time (code analysis)/Database time(sql, etc)
Rendering & Painting time
Waiting on User input
Then deep dive within the transaction, at the browser, network, application code, integration and database sql (…) levels