Re: [Hawkular-dev] Should Keycloak integration be optional?

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 01/27/2015 10:34 AM, Thomas Segismont wrote: While I agree that an optional KC would make the above points easier, I'm failing to see the concrete advantages that it would bring, specially considering that the main code will depend on features that KC is today bringing (multi tenancy security, for one). For instance, to start a development environment, one can run the start.sh script that currently takes care of installing Wildfly and Cassandra. With the PR, KC is added to this mix. About the integration tests, I'd argue that the correct way to run the integration tests is with the security framework enabled, whatever this framework is. The reason being that this framework (plus the container itself) would provide tenant information to the target code. What we have on the PR is exactly this: the tests know nothing about KC, the target code knows nothing about KC, but KC is installed on the container and provides user information to the target code (or blocks the request, if the wrong credentials are provided). For tests where no tenant-specific logic is performed, a common credential can be used for all tests, which would make more sense than disabling the security framework for those tests. I'm not sure what you mean in your last point, though. Indeed! I think discussing this at this phase is critical, so that we can get into a consensus about the expectations :-) - - Juca.