Usual way of incident detection

How are we notified that the application doesn't work?

Not being able to use the application. Getting or making angry calls. Or mails. Or both.... very depressing...

How to check what the problem is?

Logging in to the host machines one by one, opening the log file and reading it... a little bit time consuming...

and the usual way of the incident resolving

In 70% (or even more...?) of the cases one of the well-known work flows has to be executed (e.g. restarting the server, removing some file to free up space, etc.)

How Reaction can help?

Incident Detection

Patterns can be defined that are monitored in the log files. If there is a match then it means an incident occurred that has to be dealt with.

Incident Resolving

Flow of commands can be created that will be executed on one or more hosts. A command can be an operating system command, a mail sending command or an if-else condition.

Presentation

gives some insight about

what problems Reaction tries to solve
how it works
and how can we make it work

In the Case Study section some video can be found how to install and set up Reaction and how to remedy an incident with the Engine.

The presentation can be started by clicking on the play button in the middle. It can be viewed in full-screen mode by clicking on the right bottom icon, the navigation can be performed with cursor buttons or the mouse.

Presenter video

about how Reaction Engine can resolve an incident automatically

It simulates a memory leak incident in the Hermes CRM application. The administrator has to approve the flow before the start.

In the Case Study section of the presentation and in the documentation this video and others can be found.