A Deep into Performance Testing Strategies with K6

Testing the reliability and performance of your systems and catching performance regressions and problems earlier is not an easy test. Monitoring and observability tools can give you an idea of some of these performance issues, but only after it happens.

There’s the option of providing a huge amount of resources to scale your application, but there is a limit of how much you can scale vertically, and scaling horizontally can be pretty expensive.

In order to balance the costs, the resources, and the use of infrastructure for different scenarios and be prepared for moments of chaos, let’s take a deep dive into the most used types of performance testing and how Grafana K6 can help us with it.

Grafana K6

Grafana k6 or just k6 is an open-source load testing and performance testing tool. It’s part of a suite of tools from Grafana Labs, a software company that specializes in data visualization and monitoring solutions. Such as Loki, Grafana, and Mimir.

K6 is a performance test by code, which means the load test scripts are plain code to get all the benefits of version control. The tests are written in Javascript, to make the script creation easy and practical.

But K6 does not run in NodeJS. JavaScript is not generally well suited for high performance. To achieve maximum performance, the tool itself runs in Go.

A deep dive into Grafana K6 load test types

To prepare for performance risks, there are some load-testing strategies that can be used, each type, testing a different scenario, with a different goal. 5 of the most common test types:

• Load Testing
• Stress Testing
• Spike Testing 
• Soak Testing
• Breakpoint Testing

K6 Loading Testing explained

Load testing applies when the goal is to understand how your system behaves under normal and expected load conditions, like a typical day. It also helps to understand if the system meets the performance criteria.

Throughput behavior in the test:

Sample code:

What K6 Stress Testing is all about

Stress testing is a type of load testing used to determine the limits of the system. 

The purpose of this test is to verify the stability and reliability of the system under extreme conditions.

Throughput behavior in the test:

Sample code:

How K6 Spike Testing works

Spike Test is a variation of a stress test, but it does not increase the load gradually. Instead, it spikes the load by a huge number of concurrent users or transactions in a very short period of time

Throughput behavior in the test:

Soak Testing with K6

Soak testing is used to validate reliability and stability of the system over a longer period of time.

In order to do that, determine the maximum number of users your system can handle, get the 75% – 80% of that number, and set VUs to that number.

Throughput behavior in the test:

Sample code:

The K6 Breakpoint Testing functionality

The goal of Breakpoint testing is to find the system limits. Revealing if it’s possible to understand if the resource consumption is in line with the expectation and if the improvements that can be done are having an effect.

Throughput behavior in the test:

Sample code:

Final thoughts

To accomplish a good analysis of system performance, a good practice is to combine load testing with more monitoring and observability tools, such as Prometheus, Grafana, DataDog, and ELK Stack.

With a complete toolbox, it will be easier to have an understanding of an application’s behavior under different scenarios and how it can be scalable.