I have an wondering about Preview Release in Machine Learning Toolkit and Showcase Documentation.
They says, Machine Learning Toolkit App is currently designed to only run on a single search head.
(https://github.com/splunk/ml-toolkit-docs/wiki)
I use Splunk Enterprise in Search Head Clustering and Indexer Clustering environment.
I would like to know what cannot be done via Clustering environment?
We just released the GA 1.0 version of the Machine Learning Toolkit & Showcase yesterday which supported Search Head Clustering and have removed the 50,000 limit on the fit command with this release as well. Please download the latest version posted on Splunkbase and check it out. Hope this helps!
To further expand on requirements, can MLT run on a 6.4 searchhead that connects back to 6.3.5 indexers?