When you use the uptime or top command for example, you're given a load average. What exactly is a low, normal and high average? What should I be looking for exactly? I don't believe it's in percentage either.
This confuses a lot of people, and the answer isn't really ever clear.
My personal experience is that the load average needs to remain below 1.00 per CPU on the machine. I count that by physical sockets, and don't use the same ratio for DUO CPUs or dual core....
I'm sure some googling will reveal a more informed answer, but as a I say, I see things go poorly when load runs greater than 1.0 per CPU.
(i.e. 4 CPUs, load of 2.5, not too worried... probably stressing RAM or disk more than CPU).
Good luck.
D.E.R. Management - IT Project Management Consulting
It's the average number of running and scheduled processes over the last 1, 5, and 10-minute periods.
Definitely below 1, the closer to zero the better, over time. Occasional spikes are normal.
High load can be caused by cpu-intensive processes, but also by processes causing high iowait. Low memory, which causes paging, which causes iowait, can also increase the load number. When I've seen high load caused by disk congestion, usually the CPU% isn't very high.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.