Mainly, this is a set of actions for websites that are supposed to work under high load in purpose for capacity planning
and one-time campaigns when we have normal traffic peak.
In order to determine the key parameters of the system, we create special conditions that will be similar to the real-time load and will be experienced by the site, server, and other systems. Load testing does not examine the speed, but tests the behavior of the system under load.
Data reporting will be shown on a graphic chart where you can follow the dynamics of website behavior under different load, define the response time performance, check if there’s any resource overload at the same time, etc. As a result, we have the capacity indicator that describes in details the stop point of the user service process.
The most important and laborious is creating a good load profile. We compile a user model and input behavior patterns shown as script and parameters that simulates a realistic load on this system.
During the active usage of the released product, website, system or application may unexpectedly stop working.
Depending on a scale of product and its outcome orientation, it may occur dire consequences.
For example, an online store has worked for a certain amount of time with one level of website traffic but running an advertising campaing caused sudden traffic rising, so website just broke down. The inoperative website may cause a huge loss of money. Such a flow of users was really unexpected and that’s not unusual situation when surpassing expectations bring damage to business. Load testing can prevent such disappointment and see clearly the product's behavior.
What questions are recommended to be asked on the stage when the website works fast?
We highlighted the following list:
• What is the comfortable maximum number of website guests, at which the speed of page generation is still comfortable for users?
• What is the maximum number of visitors, leading to its complete incapacity?
• What will be the total reason for site break down upon rising of load? Whether it is process time, core memory exhausting, network circuit blocking, etc?
The last point is especially important if expecting a new flow of visitors (engaged by advertising campaign) since their browsers do not have cached resources from your site yet.
If the width of the network channel is not very large, the website may fail simply because of the traffic. In order to consult you regarding the issues above, QASpace team offers our services.