Walker News

How To Interpret The Linux Load Average Triplets In Top Command Output?

When talking about Linux performance, the load average triplets in the output of uptime, top, w, etc, are commonly used as one of the indicators. So, what is this Linux load average triplets?

The Linux load average triplets indicates the number of processes in run-queue waiting to be processed, plus those currently executing, over the past 1-minute, 5-minute, and 15-minutes internal.

Apparently, it is not a percentage, if you read the description again!

When the Linux load average triplets increases, it means the CPU is heavily occupied for certain tasks and blocking other processes from being executed (so more processes are lining at the queue over the time).

On a machine with single CPU, the optimal value of load average should be in the range of 1 to 2, but could be as high as 10 on multiprocessor machines. Therefore, the sar command might report high CPU idle percentage while the load average seems to be extremely “high and abnormal”.
Some experts suggest that the Linux load average can help to identify source of problem, whether is on Linux server or network infrastructure, if all of a sudden the network application performs poorly – if load average is high, it is likely the network card driver issue, otherwise the source of problem is likely on network infrastructure.

It also worth to know that some experts found limitation of this Linux load average. Dr. Neil Gunther is one of them, who studied the triplets mathematically and concluded the Linux load average limitation (see also Load Average – Not Your Average Average and Load Average Reweighed).

Custom Search

2016  •  Privacy Policy