Apr 18, 2023
Some load tests are special cases. Take one of our startup design partners. For them, the point of scale tracks with the number of Kubernetes entities connecting to and writing to the particular microservice to be tested. In fact, this is his most-important load use case. Today, his tests are on-demand and manual. He creates multiple agents, and as he scales their number he keeps his eye on IO levels, UI data population, logs, and service resources. How much work would it take for him to manually design an automated test? Can we automate this in Skyramp to make this design partner’s life easier and spare him the one-off? Should we? It’s not just about mocking connecting services and writing loads. It has to do with the sequence of configuration, timestamped entries, and other complexities. We learn the most from concrete cases. Details can reveal a common denominator and a major use case. Or they can lead to corners we’d best avoid.