Server Load refers to the average number of processes waiting to be processed by your server's CPU. The higher the average, the more processes are waiting to be processed.
In a nutshell, these numbers reflect how busy your server is. Although anything calculated can be considered a process, for a server serving websites, we can assume it is likely a PHP process. Many of the websites on the internet today run on PHP coding.
For the following example, we take a website running WordPress as a Content Management Suite (CMS). When the WordPress site is visited by a user, the server processes the PHP code into static content, HTML, and the processed site is delivered to the viewer's browser. The CPU on your server executes the PHP code.
When a server is considered to have a High Load Average, the number of processes that need to be processed by the CPU has grown to an amount that is difficult for the CPU to keep under control. The processes are effectively placed in a queue for the CPU, which causes delays and effectively slows down the speed at which your site is delivered to the end user.
Suppose, for example, that your server does not perform caching and you run a promotion on your website and receive an immediate surge in traffic visiting your site. The CPU must process all PHP code for each of those requests to visit your site. For a larger, more robust server, a traffic spike like this might not slow down how responsive your site is. But for a smaller server, this can be a problem, as it may not be powerful enough to handle all incoming requests.
In addition to an increase in traffic, inefficient PHP code can cause stability issues for your server. It is important to have a trusted developer check the code to ensure that it is written efficiently and remains up to date. By keeping your code up to date, your server remains stable, but you also ensure that any security vulnerabilities are patched. A high number of requests, inefficient coding, and a compromised CMS can all lead to a spike in server load.
You can view your CPU usage as merging lanes on a highway. The main road is your CPU, and the cars on it are processes that are already running. The cars trying to merge are new, incoming processes. If there are too many cars on the main road, the cars merging get stuck, causing delays.
Just like with the highway, the number of processes a CPU can handle depends on the number of cores the CPU has. A single-lane highway can handle less concurrent traffic than a four-lane highway. Similarly, a single-core server cannot handle as much concurrent traffic as a server with four cores. As a rule of thumb, the maximum load the server can handle is equal to the number of cores it has. A single-core server is considered "maximized" at an average load of 1.00. A server with four cores reaches a maximum at an average load of 4.00. However, higher numbers are achievable and may not indicate problems.
