UF parallelization for Splunk Cloud

[X_Kinkead]

I'm looking for advice for avoiding the 'Funnel Effect' for a small number of intermediate Universal Forwarders to about 15 cloud indexers - my best lead is to apply parallelization - has anyone had success with that, specifically to the Cloud version?  Are there alternatives?  

Here is a presentation on that issue, where the problem is described at depth but the parallelization solution is only briefly touched upon:

Video - https://www.splunk.com/en_us/resources/videos/worst-practices-and-how-to-fix-them.html
Presentation - https://conf.splunk.com/files/2017/slides/worst-practicesand-how-to-fix-them.pdf

Thanks for any advice!

[X_Kinkead]

It turns out that there was a change in corporate policy which allows for each of these systems to access the 'outside' network segment, provided a business need could be given, so I'm good to go - thanks!

[richgalloway]

My advice and (IMO) the best way to avoid the Funnel Effect is to eliminate intermediate forwarders.  Usually, they're unnecessary.  UFs should be sending directly to the cloud indexers.  The network team may complain about too many firewall rules, but that's often just laziness.