Monitoring Linux server with server stats

Question

Hi,
I am planning to test our new application which will be hosted on Linux servers. When running my performance tests, I want to monitor the key metrics on these servers. I'm thinking CPU, Disk Space and Memory Available as a starting point. Can someone help me with how to monitor the Linux servers, and if these are the right metrics, as well as additional metrics that would be appropriate to monitor during tests?

Answer 1

Jack,
one way of monitoring your Linux system is to use the Virtual Agent monitors, they allow you to hockup to your linux/unix system via Telnet/SSH and to execute commands like for example vmstat to get metrics back during your testing.
CPU, Disk and Memory are good starting points as allready mentiond in this thread...in addition you may wana add counters that are more "specific" to the server type you are running like for example counters that are usefull when running a Web server or an Application server.
usefull counters might be:
Web server
*CPU Usage
*Memory Usage
*Hits/pages per second
*Web process system utilization
*Current connections
*Disk swapping
*Disk utilization
Application Server:
*CPU Usage
*Memory Usage
*Hits/pages per second
*Request queue
*Application Server process system utilization
*Queue wait time
*Disk swapping
*Disk utilization
hope that this helps
/m