One of BMC’s partners, Aternity, wrote a great piece on the Top 10 Myths of User-Centric Monitoring. While reading through it, I realized that it wasn’t truly clear where the use of desktop agents should fit in an overall monitoring strategy. When is it necessary? What will it cover and what will it miss?
Desktop agents sit in the background on your local PC and hook into the events generated by the software applications on your PC (browsers, Citrix apps, etc) while you are working. These events come from the mouse (clicks, moves), keyboard, windows (open, close, maximize), browser (load URL, paint screen), and other peripherals or apps and they tell a very complete story. Agents are a very powerful tool in the arsenal of anyone that is looking for the holy grail of perfect end user visibility.
The benefits of using agents are fairly easy to determine:
- Indisputable insight into the actions taken by the end user and their outcome, including performance metrics
- Coverage of any application, regardless of protocol or technology
However, agents come with their own challenges as well:
- You can only use them if you control the platform – e.g., the users are your employees
In short, this is a great tool for enterprises if they are trying to monitor intranet applications. They are not as good for public-facing or extranet applications since the end user is not under your control and cannot (and should not, for legal reasons) be forced to deploy an agent.
The good news is that extranet and public-facing applications tend to be web-based and the browser vendors have been recently moving further and further towards turing the browser itself into a monitoring agent. All of the most popular browsers now support the Navigation Monitoring spec with the basic information it provides on load times. More and more monitoring tools are starting to leverage this data, although it still leaves a lot of the user market in the dark since legacy browsers do not support this spec. However, as legacy browsers die off and the browser vendors evolve their monitoring data, the browser will become the industry’s best ally in gathering reliable end-user experience data.
It’s also important to remember that the end user experience is only the start of your journey towards fixing a problem. It provides the awareness that a serious problem exists and the context around its severity and importance. However, when it comes to understanding the root cause of a problem, such a limited view can only address last-mile network issues and UI design. If there are first-mile issues or if the application itself is suffering internally, client-side monitoring solutions such as desktop agents will be blind to them.
Much like any hammer, desktop agents are VERY good at what they do and what they do is very specific. They should have a well-deserved place in your arsenal of monitoring tools. Just don’t forget the saws, screwdrivers, wrenches and other tools that are necessary as well.