Performance Testing

Understanding your product capability and scalability is crucial. Knowing the limits of your servers peek performance, as well as its overload or where and how your system can break. Identifying the bottlenecks, thresh-hold, and spike-ability, its crucial and essential to understand these as there are a number of types of performance tests that need to be conducted.

You wouldn't want a spike of users joining your service for the first time, only to have the server slow or crash, a memory leak, or log file overflow your storage space, ultimately having an outage over the weekend.

Load Testing is a reliable way to assess the performance of an application or system when multiple users are utilizing it simultaneously. This type of testing reveals how well a software product stands up under real-life based load conditions and can evaluate any flaws or issues in the application's architecture. -- Imagine creating a scenario of 2x your typical user threshold, new accounts, and using the application, but understanding how much load your server is capable of before users experience degeiargtion 

Load testing does not need to wait until the completion of a project - you can test specific components for various user demands at any stage during development. Yet, for greater accuracy in system performance and more accurate simulation of real-world user engagement, it is best to carry out load testing near the end of software development process when hundreds or even thousands of users are involved.

1. Constantly monitor the CPU usage to ensure optimal performance.
2. Troublesome disk input/output activity and inevitable delays in performance can be easily identified when analyzing spikes or sudden drops in system speed.
3. Keep an eye on the server's physical memory, so you act quickly when you need to start paging or if it runs out of available RAM.
4. Evaluate whether the service is capable of spinning up extra Kubernetes containers and shutting them down prudently to conserve cost.
5. Assessing databases to ensure that vast amounts of data are accurately categorized, verifying if clustering them together is effective and recognizing any weak or sluggish segments.

Stress Testing is technique to analyze software testing that assesses a system's stability, reliability and robustness under extreme load conditions. It aims to determine if the application can withstand unusual amounts of pressure or strain without crashing. This type of software testing also evaluates how well the product performs beyond its normal operating points in order to identify potential errors or issues when pushed to the limit.

1. Understanding your limit can help discover memory leaks when you push it to the max.
2. Utilizing multi-threading in a load test helps you uncover race conditions, like when multiple users are created simultaneously.
3. Gain insight to the Synchronization complexities that include shared cache and various timings inside the product.