Not depend on timeouts for async testing flows

[adzialocha]

Currently we use simple timeouts to wait for certain conditions to arrive before we continue with our tests (for example a schema has been built / registered, a new dynamic GraphQL API established, a document was materialized etc.). We should have a more reliable and programmatic wait to really tell that the work was done, otherwise we might end up with long waiting times or race conditions where the tests will sometimes fail.

Related PR: #266

[adzialocha]

Another pattern to use here which would not involve extending our API with test utilities is to write a helper which waits for a condition to arrive within the test, including a timeout which makes the test fail if that condition did not arrive in x seconds. Maybe there's already a crate for that ..