Ask HN: What do you use to monitor your server clusters?

7 points·Skywing··9 comments

So, my co-worker and I have been developing a tool in our free time that monitors our servers at work. Most of the existing solutions that we tried were either low quality or did not meet our exact needs.

What services or tools do you currently use to monitor your servers? I'm talking about CPU usage, memory usage, etc.

View on HN

I wrote and use this: http://yamon.klaki.net/

I wanted something as minimal as possible which still knew how to do dependencies. It's mostly a black-box monitoring tool, but it has a white-box component as well and can dump info to rrdtool for graphing history.

I'm pretty happy with it. It lacks documentation though. :-)

I use server density. Great service and worth the price. Lets you monitor almost anything and if they dont have a script to monitor what you want then you can create it in a few lines of python.

The iPhone and Andriod apps are fantastic and the main reason I went with them.

I also use server density and have been happy with them. iPhone push notifications are nice. Integrates with pager duty.

I have a very simple service on each of my servers that writes out various performance metrics every few seconds to a text file which I can then pull into graphs.

Eventually I plan to add automatic notifications when things go above or below certain levels.

At the last company I managed, we did a side-by-side comparison between CloudKick and Scout, and ultimately went with Scout.

It's not perfect, and not very extensible, but it had the singular advantage of being very reliable.

Plenty of FOSS solutions. For example nagios and munin.

We used Nagios and now switched to use Zabbix.

My servers are all Linode's and they provide an iPhone app that looks like it uses RRDTool for graphs

we use newrelic

Reply on news.ycombinator.com