– Hosted Server Monitoring

Ever find something online and you have no idea how you found out about it? Well, this is sorta the case. I think I found out about via some well placed marketing on twitter. Regardless of how I found it, I did.

There have been and probably still are a million hosted remote server monitoring services; years ago I tried most of them. All I needed at the time was a remote service to monitor my Nagios monitoring server. Lets face it–when the network connection to the server that monitors all your other servers fails, it can’t exactly send you an email. But an hour later you’ll see an email telling it is back up, implying you know it WAS down :). I tested a couple services that were basically ping monitoring… Didn’t like any of them. Either they were too expensive, overly complex or unstable. One I liked didn’t notify me a server hiccup and when I went to the site I found they were no longer in business. I finally stuck a simple Nagios configuration on a server I run for my Dad’s Office. It’s one of the servers I monitor from my network and since I give him all that free computer work, I let Nagios have a couple CPU cycles every 5 minutes to make sure it can talk back to my monitoring server.

Load Monitoring my sever via

Load Monitoring my sever via

So today I find I make a quick trip through the site and like what I see, so I decided to try it. I should mention at this point, the service is beta, but it looks like they are set to launch later this month. Any linux geek worth their salt should be able to set this up in less than five minutes. Anyone who can follow instructions can set it up in 15.

I pulled up an ssh connection to a test server I have here that doesn’t do much. AMD 64bit, couple gigs of ram, nothing fancy… it runs Xen with one virtual server running on it 24/7 (my wife’s file server) and occasionally a virtual server that is used as a second test platform for server images.  If that server goes down, my wife will tell me pretty quick, so I don’t monitor it via Nagios. Literally 3 minutes after entering my ssh password I had the script downloaded, configured, running and was seeing the updates on the website.

The agent is nothing fancy, a small python script that runs as a daemon and doesn’t even need root permissions to work (which I very much appreciate). It pushes load stats to the server every minute or so, no holes to punch in a firewall, no further configuration needed. I did script it to automagically start at boot for simplicity.

The website is very simple and polished and as you see in the screen capture above, the stats are nicely graphed for you with noted min/avg/max. There are plenty of alert options you can configure to let you know when your server is acting up, down or being slashdotted. I’m still playing with the alerts a bit, but they seem straightforward.

I will point out that I haven’t had a chance to test the Apache monitoring yet. I don’t have Apache running on that box so I’ll put it off for another day.

All in all I’m pretty impressed–looking forward to the public launch of the service and updates to the agent. It won’t replace Nagios or (wait, is there anything other than Nagios? what’s that other one called? hmm… nevermind), but for those that don’t need as many options and need a hosted solution or just a solution to monitor their Nagios box, you might give a look. I’ll keep an eye on them for a while, but I do believe I’d recommend this option to clients when it fits their needs.

This entry was posted in Tech and tagged , , , .