Cloud-native turns performance testing, tuning, and optimization on their heads. Many of the fully-managed, serverless cloud services that are leveraged have implicit scalability. These services are purchased per request and will automatically scale to meet peak and unexpected demands. For these resources, it is much less necessary to perform upfront performance testing; instead, we optimize for observability, as discussed in Chapter 7, Optimizing Observability, and continuously tune based on the information gathered from continuous testing in production. We also leverage continuous deployment to push necessary improvements. This worth-based development approach helps ensure that we are focusing our efforts on the highest value improvements.
Still, there are resources, such as some stream processors and data stores, that rely heavily on explicitly defined batch sizes...
