LibreNMS Agent

So, the plan is to implement an “official” agent for LibreNMS, to supplement, but not replace, the existing monitoring featues, and to better support applications, and sort of “consolidate” the application support we have now.

High Level Plan:

1. Make list of important / desired feature set for the agent
2. create comparison of agents based on determined criteria
3. Determine if chosen solution will meet needs, or if changes need to be made
4. Test and evaluate
5. make decision
3 Likes

Official might not be the best word for it. Maybe preferred or premiere (not a wordsmith).

But, steps look good I can’t wait to have some feedback.

Yeah, fair enough. I was confident I’d have you to give me feedback on such things :wink:

Is this in place of (or in addition to) the intention to more tightly integrate collectd/munnin?

Yeah, I think so - well, possibly. At this stage, we are looking at what we want to use as our “agent”. So possibly collectd, OR munin, but while we plan on keeping existing integrations, I think the plan is to pick ONE and focus on it.

2 Likes

Is there a reason we can’t use the same extension headers for both the agent AND snmp extensions?

extension headers? “e agent”? you mean “the agent”?

I mean can we not have the poller pick up the information for whatever is in <<>>
regardless of where it comes from?

sure, but I think the problem here is, you have to write custom code to handle that information. If you use something like collectd, it writes the rrds for you…

maybe we can also think about a possibility to get a client remote install done form the LibreNMS server?

from the server side or the web ui? Shouldn’t be too hard to store the client install script on the server and SCP it to a client, just would need to prompt for credentials.

1 Like

What is the status of your Project because o you have been looking at it for a while and wondered if you are still working on it?

Did this ever get developed or is the solution maintaining Check_MK?

Kind regards

I’m interested in this as well. check_mk_agent is great and everything but we are relying on version 1.2.6p17 from Jan 2016. The latest available from github that is in binary form is 1.4.0p39 from Jan 2019. Anything released after that is only available as source, which can be compiled of course but organizations like ours would rather not maintain building the agent before distributing it. I think they are up to version 2.2.0 now which was released a couple days ago. I love how flexible check_mk_agent is but I feel at this point we as the LibreNMS community would need to compile the agent and make it available somewhere. I also would love to see a little better documentation and available scripts/plugins for the agent released with LibreNMS.

An agent for Windows will also become pretty important. Microsoft does not natively support SNMP v3 on Windows Server and has been trying to remove SNMP from Windows Server for a while now. Speculation is that whatever is released after Server 2022 may have deprecated SNMP.

I’d LOVE to see support in LibreNMS for checking/alerting on status of Windows services and maybe even Linux daemons, perhaps with a new tab in LibreNMS of ‘Daemons’ or something? Or the processes tab could have ‘Processes’ link, ‘Daemons’ link for Linux, ‘Services’ link for Windows. Similar to how ‘Graphs’ tab has links for different types of graphs.

Sorry, lots of stuff floating around in my mind about this topic.

2 Likes

Yes, I know it is perl, but these guys make a pretty good “agent”. Maybe something worthwhile lurks in the code here?