Set requests and limits for Monitoring Console and Search Head Deployer

[BGrasnick]

Hey!

Is there any way to set requests and limits for Monitoring Console and Search Head Deployer? I could not figure it out until now and it would be crucial for our usage of the splunk-operator! Is it maybe gonna be added in a future update?

Thank you for your time & answer as well as the development of the amazing splunk-operator!

[kashok-splunk]

Hi @BGrasnick

Currently, there is no way to set requests and limits for Monitoring Console and Search Head Deployer. In the future Splunk Operator releases, we will have a separate CRD for Monitoring Console then it can be configured through its CR

[BGrasnick]

Ok cool! Thank you for the quick answer!
And what about the Search Head Deployer? Can this feature be added too? Requests and Limits are crucial sometimes so it should be configurable I think

[Akhalaka]

Separating the Deployer from the Search head cluster (I'd imagine in a way very similar to that of the ClusterMaster and the IndexCluster CRDs) would be very valuable as the resources and volumes used by the deployer likely can and should be different than those used of the cluster. It would also improve the ability to install and manage apps correctly (though this is starting to be taken care of with the AppFramework and being able to target apps to local vs shc vs idxc).

[Akhalaka]

Along these same lines, if we wanted to do something "right now", would it be an entirely bad idea to stand up the deployer and search head cluster as standalone resources (ie override the environment variables with the extraEnv and provide separate configs)? I suppose that this may not be conducive to easy scaling as the operator is populating calculated things like SPLUNK_SEARCH_HEAD_URL, but if we are in a static environment, it doesn't appear to be that bad.

I am asking for my own interest in potentially trying it this way to gain that additional granularity in config management. I know we'd lose some of the niceties of the operator, but it feels like we would have a lot more control of what happens, especially being able to manage 2 separate defaults.yaml.

[BGrasnick]

This is still an issue that I cannot configure the SH deployer separately from the SH members because I might want to have bigger SH members with a small SH deployer... Any progress here?

[pogdin]

MC scenario is delivered with MC CRD. Tracking SHCD with CSPL-1485