In order to detect mistakes in the system right at the origin and to minimize their ramifications, we rely on extremely extensive monitoring of the servers and applications we run. Ideally, we immediately spot problems and, if neccessary, are able to inform our customer about possible risks - way before consequences in performance or even availability become noticeable.
Part of any worthwile monitoring effort are, of course, all the basic parameters like server and operating system availability, CPU-load, main memory and hard disk usage. Our supervision penetrates much deeper, though: We have acquired special competencies within the field of application monitoring.
Status quo is not enough
We think it is too little to only record a site or application's status quo. Thus, we specifically hold data related to the past available. Thus, we can recognize and intercept error patterns that might have built up unnoticed and through the course of several days.
We are able to map complex scenarios, for instance by simulating click paths. Fully automatized, a test user logs into an application in regular intervals, visits a series of sites in a row and thus acts like a typical user. As soon as single sub-steps take too long - we are in on it.
Right now, we are extending our portfolio towards business process monitoring. Irrespective of a single component's state, it unveils whether the entire service or business process is available from the end user's point of view.