Server monitoring tools/solution?

Energize · 18 Oct 2015 at 16:54

I've setup a Windows Server 2012 R2 machine which functions as a database, web (lan facing), file, vpn, and client backup server.

I would like to be able to setup some kind of automatic monitoring of the server that will alert me to any major problems like low disk space, high memory/cpu utilization, raid drive failure, connection issues on the lan etc. I would also want downtime monitoring as well.

Obviously I need a remote machine to monitor for downtime.

I was wondering what the standard solutions are for this kind of thing? I assume that a remote machine has software which interfaces with software on the server and then takes action such as sending an email or sms if there is a problem with it?

Xez · 18 Oct 2015 at 17:25

We use PRTG network monitoring. If it's a Dell or HP server that it can monitor a lot of hardware values also such as raid failure, raid battery failure etc.

matthab · 18 Oct 2015 at 18:18

Depends on budget. As a start I would say nagios with mrtg plugin.

nutcase · 18 Oct 2015 at 19:02

For 1 server the 100 free sensors with prtg is ideal.

Throrik · 18 Oct 2015 at 23:11

We have used PRTG, Nagios and What's Up Gold, they've all be good at what they do, but Nagios was my favourite once it was configured. WUG was the easiest to configure.

Caged · 18 Oct 2015 at 23:18

With server monitoring you either pay with cash and buy something, or you pay with your time and rip your hair out trying to get Nagios working.

If your company has an attitude of "buy something that works and is supported" then look at PRTG. If they have a culture of tinkering then Icinga (Nagios but less horrible to look at) is worth a go.

GodAtum · 28 Oct 2015 at 09:03

Zabbix is free and meant to be very good.

swain90 · 28 Oct 2015 at 10:10

If you have any programming skills, maybe write your own application to monitor the system?

We had an issue with one of our Cloud servers where SQL would start nomming all the RAM.

Wrote a small tool in VB.net which would email me if the amount of memory available dipped below 300MB.

It's worked great for us. The only thing that I would see you might have issues with is the RAID as I'd have no idea how to monitor that, but certainly monitoring CPU and memory wouldn't be a problem.

Also on your remote machine you could get it to ping your server, if you don't get a reply, send you an email.

RichIbizaSport · 28 Oct 2015 at 19:49

GodAtum said:
Zabbix is free and meant to be very good.

Zabbix is fantastic but would be massive overkill for monitoring one server. I would definitely not advise it for this scenario.

nox_uk · 29 Oct 2015 at 10:43

RichIbizaSport said:
Zabbix is fantastic but would be massive overkill for monitoring one server. I would definitely not advise it for this scenario.

yep - complete overkill. But it is an awesome product - you can do just about everything with it, probably even monitor your toaster

Nox

Wilderbeast · 30 Oct 2015 at 09:51

Use solarwinds here, great product but not cheap, also the sales team for it are very aggressive

beavis · 30 Oct 2015 at 10:09

ManageEngine Opmanager is free for 10 devices and very easy to get up and running.

Captain Fizz · 18 Nov 2015 at 16:03

Not sure how old a nerco will be but I use http://www.eventsentry.com/downloads/full-vs-light

Happy with it, need to tailor alerts but is free and works nicely for me, email of important issues.
Added things / software I wanted checked.

rexehuk · 23 Dec 2015 at 13:24

If you want to pay... everyone raves about NewRelic.

I was going to pop in here and also say Zabbix, but agree it maybe is overkill for one server. But it would be quick to install and you can just run it as a single server/node without any hassle.

The default templates/triggers would all be there, and it would pretty much be install and go. In terms of emailing, I'd recommend tieing it to an SMTP relay like mailgun which is free.

Caged · 23 Dec 2015 at 19:41

NewRelic isn't really an infrastructure monitoring tool, it's more for looking at why your web app has suddenly crawled and lets you correlate the performance hit with a database latency issue etc.

NickTheBiz · 30 Dec 2015 at 15:26

Vote for nagios here for my requirements.

I know my private cloud hosts use Pandora so might be worth a look but it's pricey I think.

Showboat · 30 Dec 2015 at 15:30

beavis said:
ManageEngine Opmanager is free for 10 devices and very easy to get up and running.

A great product

rexehuk · 1 Jan 2016 at 16:55

NickTheBiz said:
Vote for nagios here for my requirements.

I know my private cloud hosts use Pandora so might be worth a look but it's pricey I think.

Depends on OPs experience to be honest... nagios is cool but a lot more configuration needed compared to zabbix out of the box!

NickTheBiz · 8 Jan 2016 at 10:18

rexehuk said:
Depends on OPs experience to be honest... nagios is cool but a lot more configuration needed compared to zabbix out of the box!

Quite probably, yes. I don't know zabbix myself. I'll have a google.

Boycie · 8 Jan 2016 at 13:38

Just another one to add to the mix we use Idera, previously known as CopperEgg. You pay per probe/server and there's two levels of monitoring, if you know what you are going to use in advance i.e. number of servers then you can prepay at a good discount.