Any tips for monitoring Logstash performance

(Iain) #1


I have a simple ELK stack at present which looks to be working fine. I'm running on a single machine and getting around 60 messages every 5 mins and plan to double that soon. Can some more experienced users recommend some steps I can take to monitor the performance?

I can see basic machine stats but that seems quite broad. I believe Logstash uses queues internally, can I see if I'm overloading them perhaps? Am I about to run out of Java memory or some such? do I need more workers to scale?

I send most of my traffic to Logstash via the udp plugin, so it could be ignoring a lot of traffic I may not know.

Any advice to keep my setup running smoothly is welcome.


(Mark Walkom) #2

Monitoring in LS is very basic at the moment as we don't expose anything internally, including the queue.

You could look at using the metrics filter to do something at a high level.

(system) #3