Big brother is watching: Why extensive monitoring is so important

A failure of productive systems that is not responded to in a timely manner can lead to economic damage in addition to unreliable accessibility of a service.

That's why it's important to use monitoring systems that can predict when problems might occur and perform retrospective analysis to do so. In addition, they should offer the function of comparing different data or metrics from different sources.

In order to guarantee our customers reliable accessibility of their websites and services and to be able to react to possible disruptions in good time, we rely on so-called whitebox and blackbox monitoring.

Blackbox monitoring

Here we speak of blackbox, because monitoring looks at results that are detected outside the application. In other words, how a user would see the application from the outside. Blackbox monitoring basically shows us that a service is not working at the moment, i.e. the symptoms of a service.

Whitebox monitoring

Here, the application itself provides information about the current status of its interior. The application can output much more extensive internal statistics; this makes it possible, for example, to find more precise reasons for the symptoms detected by blackbox monitoring.

In order to obtain reliable results about the reachability of services, it is essential to check them from different geographically distributed locations.

On the one hand, this can reduce unnecessary false alarms and, on the other hand, we have, for example, metrics about the loading time of a website from different locations and networks.

A great help is the configuration management tool Puppet, which we use, as well as the service discovery Consul. This allows us to ensure that new systems and services are 100% included and monitored when they are provisioned.