Slow WordPress Site? Switch to The Most Powerful WordPress Hosting

Based on the normalized requests processed within the test window, we then calculate an average possible page requests per second for the server.
This calculation shows with some guidance the theoretical page requests per second any given hosting can achieve.  Because this is a ramped load test, this number should be treated as a guidance. Many other factors can impact a server at runtime, but this number should be a realistic minimum achievable value.

We track the point in time when a server first responds with a status other than 200. 
We limit the number of error results recorded to 1 minute – as it is possible to receive requests well after the minute when a server is running slowly, but for the purposes of this test what we are looking for is when the server first starts to report errors within the minute, if it reports tests outside the minute we accept that they are beyond the bounds of the load test.

This shows the point in time within the test minute that a hosting first starts to report errors … the lower the error point that lower the number of concurrent requests the server is capable of handling reliably.

Errors arriving at the start of the test window are more severe than errors arriving later in the test window as such errors indicate that the server is less capable of handling dynamic loads. 
These sorts of early errors can cause an imbalance in the Grafna reporting system as the requests start to fail completely ( with no response ) or with an empty page type response which looks to Grafna like a valid page – but with no content

The error weight reflects pages that don’t report an error but which contain no invalid content or an empty page.

Taking all the information generated from the Grafna K6 test we create a normalized page request score for each test. This normalized score represents an estimation of successful page requests achieved within the test window, allowing for error and invalid page responses.

The higher the value the more successful page requests a server was able to process within the one minute test window.

Based on the normalized requests processed within the test window, we then calculate an average possible page requests per second for the server.
This calculation shows with some guidance the theoretical page requests per second any given hosting can achieve.  Because this is a ramped load test, this number should be treated as a guidance. Many other factors can impact a server at runtime, but this number should be a realistic minimum achievable value.