Are you confident that your app can handle a sudden surge of users? As our products start to gain traction, it’s crucial to understand how they will perform at scale. Load tests can be used to simulate multiple concurrent users accessing our applications, so we can observe how they respond.

There are a few types of Performance Testing that can be used: Load Testing, Soak Testing and Stress Testing. This article focuses on Load Testing —we will explore the different steps involved when designing and running a Load test. We will look at ramping-up load against a sample API application until just over its expected usage and observe how it responds over time.

There’s several factors that we might look out for as we perform a Load Test:

  • CPU and Memory usage across our stack
  • Response Times — mean/max/min/p90/p95
  • Benchmarking against SLAs and Non-Functional Requirements (NFRs)
  • Reliability and Stability of the application
  • Number of Errors
  • Dynamic Scaling and Elasticity
  • Performance Bottlenecks