The test data bootstrapper is reasonably stable and can be used at any level of the architecture.
As we iterate on the code, we need a way to catch breaking the system. Typically it will be if the proto or DB schema changes.
How best to automate this and not create a developer burden ?
Ideas ...
Gen API Testers...
We already gen the CLI that wrps the Proto client, and the fake test data.
We can probably generate a Test Harness that calls the CLI using the generated test data.
The trick is to validate that the results are correct using goldens. This is a hard problem because the goldens have to be created. Sometimes you can use the previous git tagged ones.
The test data bootstrapper is reasonably stable and can be used at any level of the architecture.
As we iterate on the code, we need a way to catch breaking the system. Typically it will be if the proto or DB schema changes.
How best to automate this and not create a developer burden ?
Ideas ...
https://github.com/loadimpact/k6/tree/master/samples/grpc_server