Performance: is better to deploy on host or on docker containers


#1

Hi all,
I'm new to ElasticSearch and I'm starting an analysis for using ElasticSearch both for logging and for data metrics.

I know that ES could be deployed using docker and I surely prefer this solution because I prefer to deploy an image instead of installing or update ES on the hosts.

Do you suggest to use docker to deploy ES on production with lot of data or, for performance reason, is better to install and upgrade it on the hosts?

Thank you in advance,
Marcello


(Kilynn W) #2

MarcellOr, I think it boils down to quality of deployment. Network speed, memory, cpu allocated and the config balance in the logstash instance. One thing to consider, if you go the docker route it's pretty easy to setup parallel instances and throttle resources.


#3

Yes, this is for sure but I know that performance in accessing files from the container are really poor compared to performance in accessing files from the host.

The best is to know if someone already compared both solutions with ES.


(system) #4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.