[Best practices] How many contract tests should one endpoint have? What criteria should I have to asses? complexity? stability? Let's consider phases of Contract Test intro in ecosystem: 1. getting started, onboarding new service 2. New contract in onboarded service 3. business as usual, securing service

I don’t believe it is a matter of a number of tests but instead testing what your consumer expects from that endpoint.

yeah, but there is a commonsense. what might be the rules to help me say: it is enough / too much / too few

Here is what I had done in past. In both frontend and backend code (consumer application), the code responsible for hitting provide, processing it, and then returning a response is part of util. I would look at the util and see how it is behaving against the response. If the util is checking two scenarios, one happy (200) and one bad (4xx), I would add two separate interactions -- one for each. This has been good enough as it allowed having contract tests only limited to testing instructions and not turn into provider functional test.

There is some best practice advice here: https://docs.pact.io/consumer

that is a perfect summary, thank you!