It's like trying to monitor network devices by walking an entire SNMP tree; you're going to end up grabbing much more information than you really need and cause side-effects in order to get it.
Eg. One common side-effect you may not realise is that you'll end up locking data in order to get a consistent read; or you'll end up causing a database query. You should be sure to use the whitelist and blacklist features; browse the available mbean objects using jconsole or similar to find just the bits you will be interested in, and then query only those. JMX for a production system ought to be fine; instrumentation is for the benefit of production environments after all; its also what a lot of the middleware management consoles use to expose the internals to administrator. Just don't go scraping everything; information doesn't come for free, but if you pick and choose it can come cheaply. Cheers, Cameron On Saturday, 19 October 2019 05:22:36 UTC+13, Nikolay Artamonov wrote: > > > The JMX exporter queries all of the published management beans and then >> filters the results using the expressions provided in the configuration. >> > > Hmm, at first it looks strange. I thought otherwise. May be we should open > issue? > -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/e73b9fab-b53d-4bd7-be42-9a5446c1979c%40googlegroups.com.
