Although the docs and my ./default/limits.conf say maxKBps = 0 (ie unlimited), I am receiving a INFO log entry in splunkd.conf that says:
05-26-2017 10:04:19.366 -0700 INFO ThruputProcessor - Current data throughput (258 kb/s) has reached maxKBps. As a result, data forwarding may be throttled. Consider increasing the value of maxKBps in limits.conf.
This is Splunk forwarder 6.4.2-00f5bb3fa822. I don't have any limits.conf files in my local configs at all....
I am not too concerned (we have thousands of hosts) but this doesn't seem to be the documented behavior.
The precedence (more or less) is: etc/system/local/
is GOD, then etc/system/apps/last_alphanumeric_app/local/
, then etc/system/apps/last_alphanumeric_app/default/
, then etc/system/default/
(there are some additional nuances but that's a reasonable working understanding for most purposes). So the problem is that you have an app which Splunk installs that has maxKBps=256
in it here:
etc/apps/SplunkLightForwarder/default/limits.conf
Did you mean SplunkUniversalForwarder instead of SplunkLightForwarder?
Run this:
splunk btool limits list --debug | grep maxKBps
It will tell you what the daemon sees as the true value and also tell you which file sets it. If it is set to 258 in a local/limits.conf
file then that is the file where you need to change to it to 0.
+1
btool is indeed the right way to identify the value currently in use.
As it is the default UF value, my answer was more to explain how to correctly configure it
Nope, we haven't touched limits.conf... BUT I found this:
[/opt/splunkforwarder]# find . -type f -exec grep maxKB /dev/null {} \;
Binary file ./bin/splunkd matches
./etc/system/default/limits.conf:maxKBps = 0
./etc/system/README/limits.conf.spec:maxKBps = <integer>
./etc/system/README/limits.conf.spec: pipelines will be n * maxKBps but maxKBps through each ingestion pipeline.
./etc/system/README/server.conf.spec: 1. maxKBps
./etc/apps/SplunkUniversalForwarder/default/limits.conf:maxKBps = 256
/opt/splunkforwarder/etc/apps/SplunkUniversalForwarder/default/inputs.conf file is part of the rpm and we have't touched it...
# rpm -ql splunkforwarder-6.4.2-00f5bb3fa822.x86_64 | grep limits.conf
/opt/splunkforwarder/etc/apps/SplunkUniversalForwarder/default/limits.conf
/opt/splunkforwarder/etc/system/README/limits.conf.example
/opt/splunkforwarder/etc/system/README/limits.conf.spec
/opt/splunkforwarder/etc/system/default/limits.conf
# cat /opt/splunkforwarder/etc/apps/SplunkUniversalForwarder/default/limits.conf
# Version 6.4.2
[thruput]
maxKBps = 256
w
Yes, this is the default application that comes packed with any UF.
This is expected and defines the max rate allowed.
If you do like I exposed, you will get it to work.
I have edited to remove the first line of my answer.
The good practice is creating base configuration application for Splunk, and your limits.conf must be set a local/ directory.
Hi,
Good practices:
forwarder_throughput_config/local/limits.conf
# limits.conf
# Set the max throughput allowed on UFs
[thruput]
maxKBps = 2048
And push this configuration app to your forwarders.
Cheers,
Guilhem
Notes: I have set 2048, in your case you will set "0" to have it unlimited