Having used Server Density or any Linux server, you will be aware that CPU load is not a simple percentage. Indeed, it is usually represented as a decimal "load average" over a period of time.
Server Density shows the 5 minute load average. In basic terms, a load average of 1.00 means the CPU is at 100% utilization.
Although there are many things that affect the actual value, it is acceptable for our purposes to say that this changes if you have multiple CPUs or multiple cores. For example, if you have a dual core processor, then a load average of 1.00 means 50% CPU utilization and so the 100% utilization threshold is 2.00.
A CPU at 100% utilization is not necessarily a problem. Indeed, for some uses such as batch data processing this is simply making use of all available capacity. However, this is less than ideal for other situations such as web servers. If the CPU load is at 1.50 on a single CPU web server then some processes, and hence some web requests are having to wait for CPU time. This leads to delay in responses to web visitors.
The recommended alert threshold is therefore dependent upon both the server hardware specification and the use of the server. Using the server snapshot feature you can then drill down to find out which processes are using the most CPU time and optimise your server configuration or look into adding new capacity as appropriate.
Comments