Handle polly timeout rejected exception for create engine wait

[torkins]

Please add to description a step-by-step of what Was happening to cause the intermittent failure and the new step-by-step of what will happen instead

[NRHelmi]

Please add to description a step-by-step of what Was happening to cause the intermittent failure and the new step-by-step of what will happen instead

Tests use EngineFixture to handle engine creation, the EngineFixture is shared between test classes using the ICollectionFixture the same way as it is described in xunit shared context between tests documentation

The issue is happening this way:

• One test locks the engine creation and attempts to create a new engine if _engine field from the EngineFixture class is null
• The engine creation times out and the test fails and unlocks the engine creation handler (at this level the engine is in PROVISIONING state)
• Another test locks the engine creation handler and attempts to create a new engine cause _engine field from the EngineFixture class is still null: the test calls Client.CreateEngineWaitAsync while the engine is in PROVISIONING state and fails.

• The other tests fail in the same way with this message: Conflict: engine already exists.

  Instead we edited the EngineFixture to check if the engine exists or not:

• If the engine exists and is in PROVISIONED state, the test will use this engine.
• If the engine exists and the state is not PROVISONED the test will fail with an EngineProvisionFailedException.
• If the engine doesn't exist, the test will try to create a new engine the same way it was handled before.