mgr in¿. Piotr Pêgiel

"Self-healing in the monitoring of distributed systems"



 The presentation concerns the self-healing aspects of the monitoring
of distributed systems. Nowadays, while distributed systems 
are getting increasingly complex, monitoring facilities should become 
more "intelligent" - as the first step they can guide the user what 
should be monitored. The next level of the "intelligence" can be 
described by the term "self-healing". Its goal is to provide 
the capability that decisions made automatically by the monitoring 
system should force the system under monitoring to behave more stable, 
reliable and predictable.
During the talk a new monitoring system will be presented. AgeMon, 
is an agent based, distributed monitoring system with welldefined roles 
which can be performed by the agents. So far four different roles 
are thoroughly designed and developed: GUI role, Rule role, Monitoring 
role, and Persistence role.
The self-healing in the monitoring system and the monitored one can
be achieved by the cooperation of agents performing these roles.
The presentation addresses the roles in the system together with 
the implementation background. When the self-healing of the monitoring
system is concerned, a good example is the case where it is possible
to lose monitoring data due to problems with storage. AgeMon handles
such problems and automatically elects substitute persistence agents
in order to store the data. The second aspect of the self-healing -
healing the monitored system - is also supported in AgeMon. This can
be achieved by the combination of rules and actions, executed by
the Rule role, when the monitored system enters a defined state.