Splunk or ELK to parse Bro logs

Hi all,

I would like to do some tests and deploy rules using Bro under my laptop test lab. Due to limited resources, I would like to install some log parser tool for Bro like Splunk or ELK.

In the past, I have used Splunk and goes very well for my needs. But doing some searches, I am finding more docs about using ELK with Bro than using Splunk.

But I don't see how can I limit resources assigned to an ELK infrastructure to suit my needs (I can't assign more than 2.5 GB of RAM to vm where I will install splunk or elk or another solution).

I don't expect a lot of logs per day or hour from Bro's side (in fact, I expect very few), but i don't see pretty clear what solution to use.

What are your opinions or recommendations?

Many thanks to all.

You could just change the JVM you’re using elasticsearch/logstash on to only allocate 1GB of RAM. On that VM if you give it 2.5GB of RAM, then only 1GB of it will be used by your Elasticsearch install. The rest will be used by the OS (disk cache) and logstash.

In CentOS land, you’d make your /etc/sysconfig/elasticsearch file say this:

ES_HEAP_SIZE=1g

Cheers,

JB

For Splunk I’ve had good luck with logging to JSON format and using this TA

https://github.com/jahshuah/splunk-ta-bro-json

Jeffrey Collyer

Many thanks to all for your inputs. Regarding to use ELK, is it safe to use latest versions of Logstash, Elasticsearch adn Kibana? What version do you recommend?