[RESOLVED] Predictive Dependency Checks

[xkilian]

According to the wiki, Predictive Dependency Checks are not implemented.

Premise:
Host B is dependant on Host A.

If host B goes down, will Shinken set Host B as soft-down and re-schedule Host A to resolve the dependency and not send out a erroneous notification.

Thank you for clarifying this.

X

[naparuba]

Yes, that why here it's "predictive" from Nagios point of view. It's when you launch A, you also launch B to know. But it's a killing performance for something like 3s delay for the information (and because of the internal nagios code, when Shinken got a real dep tree, when Nagios it's more immediate state).

[xkilian]

Thank you for the answer and welcome back from vacation.

This is applies to hosts and services in addition to the normal parent/child dependencies. I can understand these are not used very much and have a big impact on performance.

I will update the wiki to make this a bit clearer.

You may have noticed my ongoing cleanup of dead branches of the official documentation. There was a lot of cruft from the Nagios days that does not apply to Shinken.

Cheers,

X

[Marksmith]

Host and service dependencies can be defined to allow you greater control over when checks are executed and when notifications are sent out. As dependencies are used to control basic aspects of the monitoring process, it is crucial to ensure that status information used in the dependency logic is as up to date as possible.