Anyone who has ever tried to develop and maintain applications that react to different system events such as out-of-memory or syslog-message, will quickly be confronted with the fact that many of the available were not designed to be machine-readable. This means that even an minor update can make entire implementations unusable by changing log messages.
The Elos project aims to bridge this gap and offers a solution for converting various sources in different formats to a standardised machine-readable format and to propagate them in a Linux system.
Using selected event sources and events of a Linux system, the specially developed standardised event format is presented. Insights into the structure of Elos and the still young ecosystem of plugins are intended to show how applications can react to system events without having to know and maintain specific details of the respective Linux system.
The presentation shows an elos reference system that reacts to failing processes and sends messages with the relevant information.