Refactor the test infrastructure and develop new tests for data sources or engines

[lhubert]

Expected behavior and actual behavior.

Tests should be reworked and the framework to run them should be revisited. If we use the github infra we cannot run all tests. May be setting up our own kubernetes cluster and every night do a git pull, running the tests on our infra and provide whatever needed to alert when tests fail is the way to go.

Steps to reproduce the problem.

Specifications like the version of the project, operating system, or hardware.