You are developing a distributed microservices-based application on Kubernetes and have created alternative versions of a service. You want to identify the
best version of your service using a live experiment and rollout this version in a safe and reliable manner.
Iter8 is an open source toolkit for
continuous experimentation on Kubernetes. Iter8 enables you to deliver high-impact code changes within your microservices applications in an agile manner while eliminating the risk. Using iter8’s machine learning (ML)-driven experimentation capabilities, you can safely and rapidly orchestrate various types of live experiments, gain key insights into the behavior of your microservices, and rollout the best versions of your microservices in an automated, principled, and statistically robust manner.
iter8 experiment to safely expose alternative versions of a service to application traffic and intelligently rollout the best version of your service. Iter8’s expressive model of experimentation supports a diverse variety of experiments. The four main kinds of experiments in iter8 are as follows.
winner, where the winning version is defined by the criteria you specify in your experiments.
iter8will verify if the version meets the criteria you specify in the experiment.