Looking for feedback from actual customers on SevOne for network monitoring
. anyone using them and willing to share thoughts online/offline?
They have an appealing system for network monitoring and considering it as a
replacement to Solarwinds.
I looked at SevOne and liked the product a lot. One thing we found was that the pricing model escalates pretty rapidly because they count every OBJECT you monitor, not every device. So if I am looking at Bytes In, Bytes Out, Errors In, etc on a single interface those are all counted as a separate OBJECT against your license count. You really have to be more selective about what you want to see which to me is really inconvenient because often you don't know what SNMP object you want to look at until a problem surfaces. One of the strengths I really liked was the trending capability that helps you predict capacity issues before you hit them.
Summary: Good product, real expensive in wide deployment.
Depending on what you're after observium might be worth looking into. I run solarwinds, paessler and observium but neither are as clear and as useful for monitoring network as observium ( My opinion only of course )
Can Observium alert on SNMP traps? I seem to remember that it couldn't do
No, as far as I know that's work in progress at the moment. The alert system works well for anything polled though but depends how often you're polling
I took a look at SevOne back when you could download a free, 500-element version of it when I was looking for something to deal with Netflow. I'd heard of it prior but nothing from the website seemed overly appealing. Actually -using- the product though it was wonderful seeing a tool built to automatically deal with a lot of the things that are fairly routine but are time consuming to deal with. Automatic filtering of what is monitored based on user customizable rules. For example: Junos device? Ignore all file systems that are mounted from /dev/md*, ignore pim([de])|lsi|gre|ipip|dsc interfaces, and so on. If an interface is set to admin-down automatically prevent alarms from it. Then don't alarm on it being down. If it later changes so it isn't admin-down then start monitoring & alerting on it again automatically.
As Steven pointed out though the pricing model escalates rapidly since they do it by each individual object. If using netflow, each netflow interface is considered 100 elements if I remember correctly. Even if I ignored netflow, a single EX8216 would consume a few thousand elements or more if I wanted to monitor all of the interfaces in the chassis. Just looking at it for lab usage over ~12 Juniper devices, if I wanted to get full monitoring over all devices, without netflow/sflow, it was a few hundred thousand elements. When I try to extrapolate that to our production environment with thousands of network devices I can't even imagine what the element count and subsequent cost would be. When comparing against similar tools the cost is simply outrageous due to the licensing. And I just realized that it actually becomes more cost effective to have an internal development team dedicated to writing & maintaining custom network monitoring tools when compared to licensing costs like this.
Independent of that, I'm miffed that the free, 500-element version I was using for home and lab use is no longer usable. It says the license is valid until sometime in 2031, but won't actually let me beyond that point until I upload an updated license file. Can't even do a reinstall since the original license file is only valid for a few weeks before it expires. I keep forgetting to contact support about it when I'm at home but since they completely removed the free version I'm doubtful that they will provide an updated license file.
So yeah, fantastic tool, not as pretty as Solarwinds, but it gets really expensive, really fast. And when talking with them I got the impression that the licensing was per year versus a one-time license cost and then recurring maintenance cost for support & software updates; the above licensing behavior in the free version supports that impression. I don't know if that is correct though as I didn't think to ask while I was talking with them.
I've been using SevOne for 3 years, and I can confirm some of your suspicions around element licensing, in that you will consume more element counts than you allowed in your budget. It does provide a very granular way of omitting objects from discovery through regex. It is not a single pane of glass solution, in that fault management is not its forte. This platform is for performance measurement and management primarily. A good example of this is that out of the box, it cannot throw an alert if an interface goes down. You have to programmatically build each alert based on an SNMP polled value, so there is a long lead time before you can bring it into production. Compared to other similar products out there, price per license seems to be pretty steep, since it include the hardware, but you also will continue to pay 18% maintenance year over year.
I'm available for any one-on-one discussions you might have about the platform offline.
-The boundary to your comfort zone fades a little each time you cross it. Raise your limits by pushing them.
This electronic mail transmission may contain confidential or privileged information. If you believe that you have received this message in error, please notify the sender by reply transmission and delete the message without copying or disclosing it.