Getting Data In

Memory Usage by streamfwd.exe

ashajambagi
Communicator

Hi All,

I have recently deployed Splunk TA Stream on universal forwarder to collect DNS data. Stream App is configured on heavy Forwarder. The universal forwarder is forwarding the data to indexer cluster.

The streamfwd.exe service on DNS server is consuming 1GB of memory. Is it a normal behavior of streamfwd.exe service to use memory in GB?

UF host details : Windows 2012 R2 , Memory : 32 GB , 64bit

Below configurations on Universal Forwarder:

limits.conf

 

 

maxKbps = 4096

 

 

inputs.conf

[streamfwd://streamfwd]
splunk_stream_app_location = https://<HF_IP>:8000/en-us/custom/splunk_app_stream/
disabled = 0
stream_forwarder_id =
sslVerifyServerCert = false

 

Labels (2)
Tags (1)
0 Karma
1 Solution

Richfez
SplunkTrust
SplunkTrust

How busy is your DNS server?

Also, you've limited the maxKbps of the UF to 4 Mb.  If during busy times the DNS entries exceed 4 Mb, then it just buffers it all, and that would use a lot of memory.

If I were you, I'd raise those limits WAY up higher, or remove then completely, and see what change that makes.  Try it at 'maxKbps=0'  (Which is unlimited)

You can always set it back to something less than unlimited after testing proves this solves it or does not solve it.  Frankly, I'd just leave it set to unlimited and build out indexer ingestion if you have to.  The only reasons I can think of to leave it limited is to not fill a small pipe, like a WAN connection that's underprovisioned for what's needed.

 

Happy Splunking,

Rich

View solution in original post

0 Karma

Richfez
SplunkTrust
SplunkTrust

How busy is your DNS server?

Also, you've limited the maxKbps of the UF to 4 Mb.  If during busy times the DNS entries exceed 4 Mb, then it just buffers it all, and that would use a lot of memory.

If I were you, I'd raise those limits WAY up higher, or remove then completely, and see what change that makes.  Try it at 'maxKbps=0'  (Which is unlimited)

You can always set it back to something less than unlimited after testing proves this solves it or does not solve it.  Frankly, I'd just leave it set to unlimited and build out indexer ingestion if you have to.  The only reasons I can think of to leave it limited is to not fill a small pipe, like a WAN connection that's underprovisioned for what's needed.

 

Happy Splunking,

Rich

0 Karma
Got questions? Get answers!

Join the Splunk Community Slack to learn, troubleshoot, and make connections with fellow Splunk practitioners in real time!

Meet up IRL or virtually!

Join Splunk User Groups to connect and learn in-person by region or remotely by topic or industry.

Get Updates on the Splunk Community!

Agentic SOC Triage: Investigating Splunk ES Notables with MCP Server and a Local LLM

The Problem: Too Many Alerts, Too Little Context Security operations teams running Splunk Enterprise Security ...

All Work and No Play? Not at .conf26! Unwind at These Evening Events

Between hands-on technical sessions, keynote reveals, and diving into live architectures, .conf26 is packed ...

Join the Hackathon at .conf26 and build a No-Code AI agent

Join us for the AI Agent Buildathon, an in-person, three-hour hands-on Hackathon where you’ll use Splunk Agent ...