Looking for server/workstation/network device monitoring software (with a twist)

TSCrv

Senior member
Jul 11, 2005
568
0
0
Ok, ill try to keep this short and simple, my company's MSP dept remotely monitors and fixes issues for a number of clients, from customers without an AD structure to companies with multiple sites, vpns, etc... and we are currently looking for another hosted (or host it yourself) monitoring package with a little less problems than N-Able

Currently with n-able we have a probe on one server at a site, and every computer and server gets an agent installed. the agent checks things like cpu/ram, hd space, logs (not very well though), connectivity, services, and is a remote access agent (much like logmein on steroids) The probe on the other hand, checks things like connectivity, cpu/ram/disk via WMI, and monitors items that you cant put an agent on, like a router. (it will check ping, telnet, http, etc...). Also notifications of when things go wrong, cpu spikes, agent stops responding, etc.. Through the N-Able dashboard, we can see everything's status, and more often than not, has a lot of red Xs of things that n-able thinks is wrong with the system.

Heres an issue, with all the "problems" that may not be problems, i can go through server by server and configure them so i only get notifications if it stops responding, sql dies, etc... but with N-Able it is currently painful. I have spent 20 hours on a server to get it to purr like a kitteh and i get critical emails about something being down when i need to know about it, but i cant spend 20 hours on configuring monitoring options when i cant copy that template to another client because their structure is completely different. What im looking at is more like a few hours for setup, maybe an hour or 2 for maintenance per server every month or so (not server maintenance, maintenance on the monitoring software on the server.)

N-able is fairly expensive, im not sure what we pay for under 500 devices total across all clients, but for us to host it ourselves in house, a license for 5 years costs $40,000, not including the server. (n-able has actually complained to us because we over-tax their servers. They said we were the only ones that monitored other companies, where as normally a customer would monitor a hundred devices on their own network. i lol'ed when the tech guy said that.)

So basically we are looking for another monitoring software (hosted or otherwise), that supports multiple non-connected networks and isnt limited to just workstations and servers. Ive been a few datacenters/NOCs and have seen the monitoring alerts on a big screen in the middle, never bothered to ask what they use though, might be an option.

If anybody has names of companies / software that might fit go ahead and throw them to me. Also we arnt afraid to spend some $$; seeing how some of our contracts bring in 2-5k per month for me to sit there and check backups and remote into a user's workstation because she cant print from the terminal server. lol. Also multiple techs would be in on the action, possibly with their own dashboards or list of customers to watch.

Thanks in advance.
 

jlazzaro

Golden Member
May 6, 2004
1,743
0
0
if you consider $8000/year to be fairly expensive, run now ;P some of the major players in this market place:

HP Openview
IBM Tivoli
Level Platforms
SolarWinds
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
We currenlty use Level to monitor our customers and for the most part I like it. But with any monitoring system you're going to have to tune the alerts, there's no way around that. The templating system Level uses is good and bad, it's good because you setup one template and apply it to groups of devices. It's bad because if you have one odd device you either have to create a separate template for it or deal with false alerts.

I looked at SolarWinds briefly and while it looked really good once you added a few plugins, it was extremely pricey.
 

TSCrv

Senior member
Jul 11, 2005
568
0
0
Originally posted by: Nothinman
The templating system Level uses is good and bad, it's good because you setup one template and apply it to groups of devices. It's bad because if you have one odd device you either have to create a separate template for it or deal with false alerts..

Currently with N-able i would create a service template, and the odd ones (like the single core on a dual core template) would throw an issue, but all i would have to do is remove the service for the 2nd cpu and were good, dont have to recreate a template for it.

however, another issue with n-able, for cpu monitoring, if 1 cpu is at 93% and the other is at 5%, (49% total) it throws alerts, cant combine the 2 and run an average. (but i do see the need because if a single threaded app is pegged the way i prefer to monitor wouldnt alert me to an issue.


Originally posted by: jlazzaro
if you consider $8000/year to be fairly expensive, run now

well, (i think, only got a round number from accounting) we pay 600-700 a month right now, we charge much, much more for the monitoring. That is only monitoring about 130 devices across 6 clients, and we pay per device, with server OSs costing about 2x than xp or a router.... before I cut off 5 clients (about 1/2) because their contract expired with remote monitoring and they went to on-site stuff. so it was around $15k a few months ago.

heh, that would suck for me because my whole home network is server 2003, even my gaming boxs.


Should i even look at Nagios? it might get me a nice bonus for saving $$. I havent figured out if i can make it report back through the net to my server, ill probably find out a bit later.

My boss threw a trial of Kaseya at me ill be working with as well...

ed: ooo nice on your 23000th post
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
however, another issue with n-able, for cpu monitoring, if 1 cpu is at 93% and the other is at 5%, (49% total) it throws alerts, cant combine the 2 and run an average. (but i do see the need because if a single threaded app is pegged the way i prefer to monitor wouldnt alert me to an issue.

I believe there's a cumulative perfmon object, can you monitor that instead?

Should i even look at Nagios? it might get me a nice bonus for saving $$. I havent figured out if i can make it report back through the net to my server, ill probably find out a bit later.

I haven't seriously looked at Nagios since I used it back when it was NetSaint but back then it was a huge PITA to configure. It's awesome and incredibly flexible but unless you know unix pretty well you probably don't want to go down that path. I believe ZenOSS is another free one that's simpler to manage than Nagios, that might be worth looking into.
 

TSCrv

Senior member
Jul 11, 2005
568
0
0
Originally posted by: Nothinman
I haven't seriously looked at Nagios since I used it back when it was NetSaint but back then it was a huge PITA to configure. It's awesome and incredibly flexible but unless you know unix pretty well you probably don't want to go down that path. I believe ZenOSS is another free one that's simpler to manage than Nagios, that might be worth looking into.

*nix, well, about that.... *rolls eyes and looks away*

whenever im dealing with linux i always have a few google searches up, pretty much i can set it up, and do anything that there's already a how-to guide out there on the net. I have done some pretty crazy things with my ESX box, but probably stuff that anybody else would find normal.

I would say i can install and get all the needed services running, and if installing nagios is anything like installing apache with php, throw in mysql, and a few other things, then i might be fine.

If i were to throw it up quickly over the weekend just to monitor 2 or 3 things just to get a feel for it think ubuntu would suit?
 

jlazzaro

Golden Member
May 6, 2004
1,743
0
0
Originally posted by: Nothinman
I believe ZenOSS is another free one that's simpler to manage than Nagios, that might be worth looking into.
i would say Zenoss has a much steeper learning curve. I tried really hard to deploy it instead of a commercial solution, but I eventually gave up on it. Advanced knowledge of Linux as well as programming (more specially Python) is a requirement.

if you have all day to babysit the application and tweak code, its great. There are a few posts on the Zenoss forums on why it just isn't ready for production environments. from your posts, it sounds like you need a solution that just WORKS. with all your frustrations with N-Able, i think an unsupported open source *nix solution will just add to your headache.

no harm is loading it up and trying it out though...Ubuntu would do just fine. if you're curious about Zenoss, they have a prebuilt Virtual Appliance for download that you can get up and running in minutes.
 

Brovane

Diamond Member
Dec 18, 2001
5,888
2,197
136
I run SolarWinds Orion NPM and it basically does everything that you want it do and then some and it is very easy to configure and setup alerts. It is probably out of your price range but it works. I would check out and see how much it would cost for you. I basically don't know how I managed the equipment without it.

I monitor over 6000+ elements without any issue. I track everything from interface utilization, CPU queue lengths, temperatures, service tags, chassis serial numbers, Disk queue lengths, CPU utilization, memory utilization, POE power draw, to what services are running, I can even track if specific websites are up and running (I.E. I get a alert if a web-page doesn't display a specific text-which means it is down). All this data is logged to a SQL database so I can go back and run reports on the data to my hearts content.
 

TSCrv

Senior member
Jul 11, 2005
568
0
0
TTT

Nagios is a no-go... from getting it setup and stuff like that, ive noticed that the server queries the client, instead of the client checking into a server. The latter is pretty much the only way to go because of me managing multiple networks in different geographic locations.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Nagios is a no-go... from getting it setup and stuff like that, ive noticed that the server queries the client, instead of the client checking into a server. The latter is pretty much the only way to go because of me managing multiple networks in different geographic locations.

That's the way most of them work because you need the server to at least ping the client to make sure the device is still up and on the network.
 

nervegrind3r

Lifer
Jul 12, 2004
16,267
5
81
back up, im in the same boat.

We currently use level platforms (hosted) at work, and I personally think its a bad product. I constantly get emails about devices going down (switches usually) that really never went down, or entire sites going down. Sometimes I will get a flood of hundreds of emails saying every device at a site went down, its just very annoying. We cannot configure the scan intervals for devices either. I am currently looking to recommend a new product that can monitor different network devices and servers via snmp and wmi (no client software can be installed), and for several independent sites. I have heard of solar winds, but have not tried it yet. Any recommendations
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
I constantly get emails about devices going down (switches usually) that really never went down, or entire sites going down.

Then you need to find the root of the problems because Level isn't it and you'll almost assuredly have the same issue with other software.

I have heard of solar winds, but have not tried it yet. Any recommendations

Orion and it's modules look nice, I only used the trial though, but the price is pretty high.
 

kavi123

Junior Member
Jan 7, 2009
1
0
0
We currenlty use Level to monitor our customers and for the most part I like it. But with any monitoring system you're going to have to tune the alerts, there's no way around that. The templating system Level uses is good and bad, it's good because you setup one template and apply it to groups of devices. It's bad because if you have one odd device you either have to create a separate template for it or deal with false alerts.

I looked at SolarWinds briefly and while it looked really good once you added a few plugins, it was extremely pricey.

===================================
IT Hardware | IT Solutions
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Originally posted by: nervegrind3r
back up, im in the same boat.

We currently use level platforms (hosted) at work, and I personally think its a bad product. I constantly get emails about devices going down (switches usually) that really never went down, or entire sites going down. Sometimes I will get a flood of hundreds of emails saying every device at a site went down, its just very annoying. We cannot configure the scan intervals for devices either. I am currently looking to recommend a new product that can monitor different network devices and servers via snmp and wmi (no client software can be installed), and for several independent sites. I have heard of solar winds, but have not tried it yet. Any recommendations

It's not the software. You have an underlying problem that needs fixing.
 

Cooky

Golden Member
Apr 2, 2002
1,408
0
76
I have to admit I haven't read through your entire post (late night), but we use HP OpenView for health monitoring and NetQos for network/apps performance.