Rapid microservices development with Kubernetes

If you’ve ever developed services to be deployed and run inside Kubernetes, you’re probably familiar with the associated pains. Each service’s upstream and downstream dependencies (be it databases or other services) makes it hard to develop locally and test in isolation.

Using a shared environment — staging or testing — sounds promising, but it turns out it’s not very easy, mainly to keep it up to date and as close as possible to production. The fact that multiple developers play in the same sandbox makes nothing easier.

Typical inner loop

If you think about the steps you go through, your inner loop or development workflow, when developing a new service, they typically include the following:

1. You write some code, make changes.
2. You build the code, including compiling and building your code and the container image.
3. You push the container image to a registry (or load it into a local registry if using a local cluster).
4. You update the…