Nagios Support Forum

Does changing the BPI paramater "XML Cache Threshold" affect the delay in how long it takes BPI to calculate that a group has changed state (and thus, alert)? And I'm assuming that this value is in seconds, since the PHP code is using filemtime, but there is no such indication in the documentation or help text. I'd suggest adding that.

Basically, we have a customer that thinks their BPI group logic takes too long to notify them when something is amiss. I'm wondering if changing this setting from the default of 90 seconds to something like 30 seconds would fix that?

From looking at the code it does look like that is the case, I'll create a documentation task for it.

Actually, it does seem to talk about it here:

bpi.xml

When service checks are run, the first time a BPI service check is run the BPI Group states are dumped to the bpi.xml file. After that
the service checks check the file age of the XML file, and if the age threshold is exceeded, a new file will be made. Otherwise, the
check plugin will used the cached results in the XML file. This file can also be used as a data API for use with external applications.

I did read that, but it's not clear how often the BPI status itself was updated, just the XML file. So I'm guessing that you're saying that if at time zero, two service checks go bad, and they're part of a BPI group, that the XML Cache Threshold value determines how many seconds until the BPI group goes bad, too?

The way the help is written, it is not clear to me that that is the case, and that it is possible that the XML file is simply updated every 90 seconds by default, but the BPI status update could be some random number, or it could be instantaneous.

I've created the documentation task with a link back to here:

The following KB article was created that explains the XML Cache Threshold in detail:

https://support.nagios.com/kb/article.php?id=568