Techniques for writing reliable API tests and QA automation

I've been building a lot of API and web app testing automation and wrote an article to share what I've learned.

To sum up some bigger points:

- use as little end-to-end tests as possible

- when end-to-end tests uncover bugs, write new unit tests

- dilligently remove flaky tests and tests that don't provide value

- keep test data consistent, short-lived, and as small as possible

Would love to discuss problems anyone else has encountered testing APIs and web services.

(note: I am a co-founder @ Assertible)