|
Load Averages Explained
by
05 Jun 2006
All "load average" indicates is a rough idea of the number of queued processes waiting for a resource plus active (currently running) processes. When you see numbers like 1.43 1.47 3.45 it indicates the load average over 3 different periods (1m, 5m, 15m) Its important to note that "load average" doesnt accurately depict how busy your system is, or why your system is busy. You could have 1 process running and nothing queued, and have 100% CPU useage, or you could be running a backup with 100% HD/Network useage. That would have a knockon effect to all the other processes in a potential queue. On the other hand you could have 20 active and queued apache threads and be at 1% cpu load. The number of cores/CPU's etc makes no difference to the load average (which is obvious from the above explanation. What does load average actually tell you then? If it's consistently very high it means one or some of your resources is probably overloaded. A very *general* idea of a good load average is <3.00. However it depends on many variables. Hope this helps a few people and isn't too technical. Let me know if you want a less techy version 
|