can you please explain why it's Ideally to install it on the HF ? and why to avoid installing it on the Indexer? and what do you mean by "unless its all in one"?
as far as I know HF is busy in parsing the data,then I'm just asking myself why HF is the better place?
in addition, where Should I configure the Inputs( input for Activity Logs/Diagnostics Logs) in splunk? is it in the search head?
Its fully dependent on your environment.
In my case we have search heads loaded with so many scheduled searches so I could not allocate even 1 cpu for modular/scripted inputs and we have Indexers are busy in responding to searches and indexing data. Thats y I recommend to have modular inputs on HF.