Hello,
I’m doing some digging on low cost solutions for the MSP I work for, and I was wondering if anyone here has used this in this way? Right now we’re using PRTG and it works well enough, but we’re digging into other options.
Can you use this for different clients across different geographical regions and have it sent to a centralized location? Can that data be kept separate (multi-tenant)? Are there any licensing issues with using it in this way?
Answers much appreciated
Thanks!
I’m an MSP and I can say that this does not work well for a multi-tenant setup. We still do use it for our larger clients but that’s just to augment our main RMM.
If i had all the time in the world, I wouldn’t mind trying. While it won’t be a true multi-tenant setup, you could separate devices by location(groups). I’m sure you don’t have vpn tunnels to all your clients, so polling will need to be via some sort of proxy or gateway. Or if you want to get creative maybe something like zerotier.
Anyway, best of luck. Keep us up to date with your research.
I personally feel that Libre is currently one of the best NMS solutions if you don’t have any budget for monitoring tools and your network is not of the hyperscale type. That being said, there are definitely feature and scaling limitations compared to commercial NMS products.
We are a specialized MSP and have used a single Libre server (along with other tools) for the past 2yrs to monitor our global network/server infrastructure consisting of between 500 - 1000 devices and 30k - 40k ports using the default 5-min polling. We run Libre on a big expensive AWS instance, and I feel like we are near the limit of what Libre can do on a single server setup at 5-min polling. While it’s possible to scale Libre by moving to a distributed setup, it also adds a lot of overall complexity that will require much more regular care and feeding then a single-server Libre setup. We would definitely need to move to a distributed setup if we wanted 1min polling in Libre.
Just to compare, one commercial NMS I’ve used in the past (Statseeker) was able to poll around 1million interfaces at 1-min intervals using a single beefy server. It licensing also cost tens of thousands of dollars /yr, so there’s your tradeoff.
In my opinion, Libre is not a good multi-tenant solution if you want to present monitoring data/alerts directly to customers. Libre has no concept of port groups, so for example you cannot create a group of all customer’s ports and limit access in this manner. Alert rules can get messy quickly too if you need to create lots of different alert rules for different customers. We monitor infrastructure for two different business units in our Libre instance and even with just this the monitoring rules are numerous and messy due to the different alerting requirements and devices of the two BUs.
Hope this helps you with your decision!
@jp-asdf, I had a thought not long ago regarding the sharing of LibreNMS data. I don’t think that it’s all too uncommon for people to want to provide a “dashboard” of sorts to different units within the same IT organization. For example, suppose your organization runs LibreNMS and you’ve got network equipment in there but you also want to use it for servers. Maybe you want to allow the system administrators some access to LibreNMS but you don’t necessarily want them seeing anything on the network side.
In order for this to work in LibreNMS natively, we’d need at least another table within the DB that takes device ID’s and lists them out against UserID’s. I don’t think this is anything that anyone who is currently developing LibreNMS is prioritizing.
But what if…using the API, we had something like Grafana where, one could create a group called “servers” and query on that and display the data? Just a thought – and doing this with LibreNMS would likely be considerably cheaper than licensing a commercial NMS.