How to configure Load Balancing on Splunk Search Heads?

[agentsofshield]

Hi! So I set up a F5 Load Balancer and listed all of my Splunk search heads as pool members.

Apparently the load balancer performs a health check, and therefore, requires a health monitor URI and a health monitor response!

So I'm consulting you guys, which URI and response should I use? It's just a simple request and response to check if my search head is up. With the default configurations my server is considered down, of course.

I have no experience with load balancers so please be gentle.

[jkat54]

I recommend monitoring the Splunkd port of 8089 (tls), because when it isnt responding it might still accept requests on the web port, but not be able to service them.

[codebuilder]

This is older but looks like it never got a solid answer. The easiest solution is to perform a simple port check.

I don't use F5, but the haproxy equivalent is such:

balance source
server shca101 :8000 weight 1 maxconn 2000 check port 8000 ssl verify none
server shca102 :8000 weight 1 maxconn 2000 check port 8000 ssl verify none
server shca103 :8000 weight 1 maxconn 2000 check port 8000 ssl verify none

Before sending traffic to any of the search heads the load balancer verifies that port 8000 is active. Assuming your web UI is using the default port 8000 of course.

[joesrepsolc]

Struggling with this same scenario... and still have no solid answer. Anyone have more info?

Thanks.

Joe

[renjith_nair]

Hi @agentsofshield ,

The simplest method is to configure an http monitor to check the service.

Select http as the health monitor and set GET /\r\n as your send string in properties. This is very basic check but should work

[agentsofshield]

I don't have a send string, I can only fill out a URI and a response.

The team in charge of servers / load balancers has a system so I have little control over it.

[renjith_nair]

Ok. In URI you could mention / which is the root context and in response you can leave blank.