Network Monitoring

D'Agger · 11-08-2015 12:41pm #1

So I've had a few threads on here regarding nagios, log monitoring, syslogs etc. & thankfully I've gotten some good feedback

It's something that interests me hugely in my work as a sysadmin - saving time, setting up metrics and organising rafts of data & information into easy to read, easy to evaluate repositories.

I've tested with Nagios, Observium, Lansweeper, Solarwinds etc. and I'd like to just get a discussion going on what setups fellow IT pros (or home users for that matter) use for monitoring, what's your setup, what do you monitor most & what do you recommend?

To Kick off...

Lansweeper:

This is very cheap in the grand scheme of things and is something that I'd highly recommend. It's very easy to setup & upgrade, runs on windows server which is helpful if you're not comfortable with linux. Gives a very nice GUI to navigate the info being pulled in by SNMP, allows running of reports, license compliance (which needs refining) but overall I'd highly recommend this for out of the box inventory management for Server, printer, switch, esx, laptop etc.

The custom fields are also very handy & can have reports ran on them i.e. add a checkbox for Documented or Licenses to servers and run reports based on that to see where you may be falling down in certain areas.

Observium:

Linux setup so will require some config before you can start playing with it. Think this is great for graphing & general look - it's very sharp looking, however the alerting is a bit up in the air for free users, I've been told it can be setup on the free license but I'll most likely need to upgrade as I'm not seeing it. Where it falls down is detail in comparison to....

Nagios/ZenOSS:

Nagios is clearly a powerful & very useful tool, however, from my own testing, it requires the most time to config & get running & even then I still don't have everything working.

The graphs aren't as good as Observium, however the map is far better, you can monitor webpages & it's up/down status monitoring is on the button with close on zero false positives so far.

I am having issues pulling in SNMP data from ESX servers & this will require some looking into but I'm at a stage now where I need to pick one solution and bury myself into it for a few days to get everything setup, however I'm still stuck between looking at other solutions (hence my interest to see what some here are running with) & choosing between Nagios & Observium - for now Nagios wins based on efficiency over looks.

Finally - apologies for the images - not many stock photos online, will look to maybe upload some of my own if we get a conversation going.

Enigma IE · 21-08-2015 12:19pm

I've mostly used/deployed PRTG but also WUG. PRTG is relatively easy to setup and has good features, WUG there's a little more to it but excellent product.

Both can monitor hundreds of different sensors and can monitor in various ways e.g. WMI, SNMP, PING, HTTP and so on. You can report on all the sensor types, create graphs etc.

Both can receive flow data (neflow, sflow, jflow) and again this be graphed and you can drill down as required.

You can also import MIB files (if available) from any devices you want to monitor. e.g. firewalls and create monitors for anything contained in the MIB file.

Personally I like to graph port and/or network link usage and identify spikes or bottlenecks.

roast · 08-09-2015 12:15pm

I'd like to give Nagios a go at some stage, might play around with it over the weekend. Out of curiosity, what was your issue with monitoring ESX? Did you get it sorted?

I've been playing around with OpenNMS a little bit recently. It's certainly hugely customizable but so far I haven't gotten it to do exactly what I want. With that said, I haven't spent nearly enough time with it yet. The MIB tools are a bit of a pain in the hoop to use in comparison to some others, but again, I probably need to sit down with it a bit more. Anyone here using it?

Standard Toaster · 08-09-2015 12:24pm

If I can help at all I'm quite familiar with Opsview, which uses Nagios as it's monitoring engine.

D'Agger · 08-09-2015 2:09pm

roast wrote: »

I'd like to give Nagios a go at some stage, might play around with it over the weekend. Out of curiosity, what was your issue with monitoring ESX? Did you get it sorted?

Have you given Nagios a lash since?

The ESX issue was that it's not collecting snmp data correctly - so I can monitor up/down status which is all I want for now really, but when I get priority of servers to be monitored sorted / set down by mgmt, then I'll have to take a closer look. I'd imagine it's something along the lines of the incorrect template being used to monitor.

I'll be looking into monitoring databases also, seeing if it's possible to monitor aspects of Oracle DBs and report on issues such as tablespace/listners etc.

Zenoss5 is looking really impressive, for anybody here who hasn't checked it out yet

https://www.youtube.com/watch?v=x8TBwXUtW44

dunworth1 · 10-09-2015 9:48pm

i use spiceworks
http://www.spiceworks.com

you can setup a central server and and other servers report back to it
and setup email alerts etc.. all free

D'Agger · 10-09-2015 10:48pm

dunworth1 wrote: »

i use spiceworks
http://www.spiceworks.com

you can setup a central server and and other servers report back to it
and setup email alerts etc.. all free

Had Spiceworks running locally on my laptop when I first started due to now ticketing system being used! Atlassain JIRA was then implemented company wide, so while everybody was on that platform it made sense to open a project for tech support & get tickets in that way. Been using it since, has it's faults but overall it's pretty handy & there's scope to do quite a lot with it

BobbyT28 · 02-10-2015 12:40pm

HI,

Any free/paid Network monitoring recommendations for how users?

Free preferably!

Regards,
BobbyT28

Enigma IE · 02-10-2015 2:18pm

PRTG freeware, 100 sensors.

https://shop.paessler.com/shop/free_license/?showkey=1&download=1

D'Agger · 02-10-2015 5:06pm

Observium for look and feel but it doesn't have email alerting on the free version so far as I'm aware

irishbucsfan · 03-10-2015 5:14pm

We use Nagios and LibreNMS.

Nagios is obviously deeply flawed and is about 5 years overdue for a replacement but it's such difficult project and the community is so splintered it's difficult to work out which way to go. Unfortunately it plays such a crucial role that big players are more likely to roll their own replacement than fork anything. Saying that, I'm looking to give Shinken a test run and possibly contribute to that, it solves some of the issues.

LibreNMS is something we've only just started integrating as we move away from Observium, it so far seems to be great but I'm going to wait until at least the end of the year before making any permanent prognosis.

D'Agger · 04-10-2015 10:41pm

irishbucsfan wrote: »

We use Nagios and LibreNMS.

Nagios is obviously deeply flawed and is about 5 years overdue for a replacement but it's such difficult project and the community is so splintered it's difficult to work out which way to go. Unfortunately it plays such a crucial role that big players are more likely to roll their own replacement than fork anything. Saying that, I'm looking to give Shinken a test run and possibly contribute to that, it solves some of the issues.

LibreNMS is something we've only just started integrating as we move away from Observium, it so far seems to be great but I'm going to wait until at least the end of the year before making any permanent prognosis.

Have you taken a look at Zenoss 5, looks to have gotten a decent facelift from what little I've seen, been hectic at work the last month so monitoring has gone by the wayside due to time required to research & config

Are you on free Observium, find it's graphing to be very good but don't have the paid version so not sure how good alerting is on it

From first glance at LibreNMS it looks like it just sits on Observium & allows alerting - are there other benefits such as reporting, I'd like to be able to pull together a report on CPU utilization for a week & get it Friday for certain servers

irishbucsfan · 04-10-2015 11:30pm

D'Agger wrote: »

Have you taken a look at Zenoss 5, looks to have gotten a decent facelift from what little I've seen, been hectic at work the last month so monitoring has gone by the wayside due to time required to research & config

Are you on free Observium, find it's graphing to be very good but don't have the paid version so not sure how good alerting is on it

From first glance at LibreNMS it looks like it just sits on Observium & allows alerting - are there other benefits such as reporting, I'd like to be able to pull together a report on CPU utilization for a week & get it Friday for certain servers

Zenoss doesn't really fix the issues that exist with Nagios as far as I'm concerned. On our network the main issue we've encountered is scalability, and I don't see how Zenoss is any better at that than Nagios. The reason I'm looking at Shinken now is that it at least distributes/shards it's core functionality onto distributed processes and allows us to scale it horizontally. I don't really care about how they look to be honest because I can just write a front end for them if I need to (so long as they have a decent API, which actually Nagios does not.... but that's probably enough for an entire thread)

We have the fully featured version of Observium, but as I said we're moving off it now to LibreNMS, which is an improvement as well as being free. It doesn't sit on Observium, it's a standalone snmpd monitoring system. You could fairly simply get reports out of it, there's a tonne of plugins available for that sort of fuctionality, we use the Graylog plugin for example which is quite nice.

It's early days with both those projects, a lot of people have been looking for a suitable replacement for Nagios for a long time but it's a very difficult task, Andy Sykes' talk on it probably underlines the issues with it better than anyone else has done

Network Monitoring

Comments