topic Splunk UF Deployment - Possible Issues in Getting Data In

Splunk UF Deployment - Possible Issues

johann2017 — Fri, 15 Nov 2019 16:28:10 GMT

Hello. We are planning on deploying UFs across our enterprise ~ 3000 systems. Currently, we have deployed UFs to 50 systems and have seen no issues. Before doing a large deployment to cover our entire enterprise - I was curious if anyone has seen any issues arise from deploying UFs?

Re: Splunk UF Deployment - Possible Issues

richgalloway — Fri, 15 Nov 2019 17:39:36 GMT

A single deployment server can easily handle that many UFs. Just remember to set phoneHomeIntervalInSecs to 600 or so (default is 60 seconds) in deploymentclient.conf.

Re: Splunk UF Deployment - Possible Issues

johann2017 — Sat, 16 Nov 2019 17:26:53 GMT

Thanks rich! Have you seen any other issues arise on the network, endpoints, servers or anything else or "gotchas" from mass deploying the UFs?

Re: Splunk UF Deployment - Possible Issues

burwell — Sat, 16 Nov 2019 20:05:41 GMT

If you are able to monitor the UF _internal log, you can possibly detect issues. The classic issue is that they maxthruput throttles the UF and the UF will get their data indexed in burtsts. If they roll logs they might miss events.

Re: Splunk UF Deployment - Possible Issues

woodcock — Sun, 17 Nov 2019 01:08:32 GMT

The rule of thumb is 1KUFs/1-minute of phone-home with default settings. The default is 1-minutes so 1K servers with default settings before you run into scaling problems.

Re: Splunk UF Deployment - Possible Issues

richgalloway — Tue, 19 Nov 2019 12:40:54 GMT

In general, no. UFs are very straightforward to deploy. Problems are always possible, however, if you exceed the capacity of any of your resources.