Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a series of virtual servers. I'm running a command to login to each one and take a look at the load averages using uptime.

What's the best way to work out if load values represent overloading? I'm running on rackspace cloud, so the servers have burst capability and can be all different sizes.

I'm a little stumped on how to come up with a consistent way of figuring out when I need to spin up new servers. I can do things like estimate the jobs running on each one, but I'd like a system that runs a little closer to the real resource use available on each instance, as it obviously varies quite a bit!

Help greatly appreciated!

share|improve this question
You can't know if a server is overloaded if you don't know what a normal load is. Have you performed any baseline performance measurements on the server? – joeqwerty Jul 8 '10 at 2:37

The Linux CPU load average tells you just that, check out this web tutorial on load averages.

A good way to keep track of your resources ( like CPU levels) across multiple computers is using SNMP (Simple Network Management Protocol) which will give you pretty much any statistic you want and is automatically updated.

share|improve this answer

Which virtualization solution are you using ? Most of enterprise virtualization solutions do have their own reporting, management and capacity planning tools like :

share|improve this answer
This. Assuming you have the relevant guest additions installed (dependant on your flavour of virtualization), you should be looking at the host's reported load on each individual VM. Don't look inside the VM itself, because RAM, CPU and IO utilisation as perceived by a virtual guest are not necessarily accurate. – Chris Thorpe Jul 8 '10 at 7:18

I use Cacti to monitor and graph server load (amongst other things).

For most servers I use 'better cacti templates' which can connect over SSH and collect data on load, cpu usage, memory, disk io, etc.

Watch out for spikes in your cacti graphs and have a trawl through your logs to find out what was happening at that time to cause the spike.

I also use Nagios for status alerts via email and SMS.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.