End-2-end measurements monitor business processes from the end users' viewpoint. End-to-end monitoring simulates typical use scenarios, measuring application response times. It should also monitor all components involved, so as to be able to identify potential problems quickly. The recorded values not only allow for easier error analysis, they can also be used as proof, if need be, detailing a specific period of time.
End-to-end monitoring is an excellent example of how open source software can be used to provide business-grade functionality that goes far beyond the standard functionality of Nagios.
An often heard phrase in companies is: "The Internet/application x is slow today" What sounds so simple from an end-user's point of view is often extremely difficult to understand.
Mr Smith has a desk full of orders that he needs to process. It is one o'clock in the afternoon and the order processing system seems extremely slow, just when he needs it most. Because he is currently under a lot of pressure, Mr Smith grudgingly accepts the fact that his work will take longer than usual. Around five o'clock Mr Smith has finished processing all the orders. He decides to go up to the IT department and ask them why the order processing system has been so slow.
A member of the IT staff remotely logs in to Mr Smith's computer and together they check the order process, which, by now, is functioning perfectly normal again. Although he is at first satisfied that the problem seems to have been resolved, Mr Smith leaves feeling slightly uneasy that the same thing could happen again tomorrow.
The IT administrator isn't entirely happy either. All he has to go on is Mr Smith's feeling that the system had been slower than usual. He now has to wade through log files and status and performance values in search of the cause of the problem. Ideally, his systems and network management tool will turn up something that is out of order and could have caused the slow performance.
As a first step, Mr Smith's subjective perception need to be captured in measurable terms. To be able to look back in time, the one o'clock data must be still available at five o'clock For analyzing purposes and in order to identify trends, the system should also be able to display graphs.
End-2-end robots perform typical user actions in configurable intervals. In addition, freely definable measuring points for various points in time allow for long-term storage of values. These can then be visualized as shown in the following graphic.
Because the systems are only used for the specific task of simulating typical user actions, it becomes much easier to rule out certain factors as possible causes.
If the end-2-end robot does not show any abnormalities, it is safe to assume that the problem lies with the user's workstation.
By using more than one robot (preferably at different locations) the additional information will reveal whether the problem is caused by a faulty network connection or by the application itself.
These examples clearly show the power of open source software. Tools like these will provide business-grade functionality that goes far beyond the standard functionality of Nagios.