Stability monitors are unreliable and must be removed in the future.

Simplest usecase to demonstrate the problem is:

1) user code is going to install service A

2) user code associates StabilityMonitor with service builder for Service A

3) user code associates Service A with service builder

4) user code executes service builder install() method

5) user code calls StabilityMonitor.awaitStability() to wait for Service A (to let it achieve stability)

6) blocking function call 5) succedes and returns control to user code

After step 6) user code might expect Service A is in stable state.

But since StabilityMonitor doesn't block services in any way this might not be true.

Code relying on it is really error prone.