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 python framework that I've created to run performance tests in a mixed server environment of Windows, Linux and Solaris. I'm looking for a way to remotely gather performance data for the CPU, Memory, Disk and Network I/O.

I can't use perfmon because the python scripts are usually run from a *nix machine. Is there something similar to perfmon that can be used on multiple OS's to gather information remotely. If it comes down to it, multiple single OS monitoring programs would be acceptable.

share|improve this question

Don't overlook good old SNMP. If you add SNMP Informant (an SNMP subagent) to the standard Win32 SNMP Agent, you can get bazillions of performance stats. Over on *nix, use NET-SNMP and hit the Host MIB.

Having done that, you can pick any of the open source monitoring tools. MRTG is a pretty easy to throw up quickly, but Big Brother, Cacti, and a bunch of others are out there. I'd avoid Nagios if you're in a hurry. OpenNMS is a nice long term solution, which I prefer over Nagios, but also not a quick install.


share|improve this answer

BigBrother does the basics you seem to need if you're after test syste-based scripted solution.

This being said, what you describe is well withing default capabilities of SNMP agent, so maybe that's the direction you should consider. SNMP is very easy to use and the beauty is once you get SNMP agents sorted on the observed systems, it will be really easy to gather stats from as many systems as you like with literally one-liner commands.

share|improve this answer

You can have a look at Xymon (hobbitmon)

It's an opensource rewrite of Bigbrother, I use it to monitor our server farm (linux) and i'm very satisfied with it. In my previous job I replaced Nagios with it to monitor linux, VMS, hpux, tru64 & Windows NT/2000 servers.

Server runs on (almost) any Unix, there is a simple & nice web interface integrated. Clients runs on Unix, Windows & VMS ( ;) ), it monitors & graphs a lot of metrics out-of-the-box (ie.: CPU, Memory, Disk and Network I/O) . You can add custom checks & metrics. It's easy to configure (and very fast). I will not describe here all features of this tool but I highly recommend it !

Homepage & demo :

Windows client : h**p://

(Sorry for bad link but I can post only one as I'm a new user ^^)

share|improve this answer
I was just reading an old page on hobbit when I saw this reply come in, Definitely looking down this path. – crackity_jones Jun 10 '09 at 15:50

Have you thought about Nagios? It can do remote monitoring of performance related metrics as you are asking for.

share|improve this answer
Nagios appears to be a lot more than what I need. I need small bursts of monitoring with a lot more detail than I'd have time to setup with Nagios, thanks though. – crackity_jones Jun 10 '09 at 15:51

The free HP solution for SNMP, HP Insight Manager programs, etc. may be useful for you - HP Systems Insight Manager - HP SYSTEMS INSIGHT MANAGER

share|improve this answer

I would recommend Ganglia. We've been using to monitor performance on our HPC and it works great.

share|improve this answer

For multi-platform monitoring a good choice is IBM Tivoli Monitoring 6.2, however this is a costly investment. IBM products are very scalable but that comes with a pretty hefty price tag.

share|improve this answer

You can also look at Tivoli Monitoring software. Its reasonably inexpensive and very easy to use. It can be used in house or remotely and you can manage your whole environment using this software from a single dashboard.

share|improve this answer

If you want something quick, on the Solaris and Linux hosts you could enable collection of data with sar(1). Basically the text equivalent of Perfmon. Then set Perfmon going on the Windows host.

Solaris installs with commented-out sar cron jobs in the 'sys' user crontab, and I think some of the Linuxes do likewise. Uncomment these to start it off.

There are tools out there, such as ksar, that can take sar output and turn it into pretty graphs.

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.