[Bug]: etcd cluster storage delete drops its success result

System Info

Platform-independent etcd cluster-storage behavior on current main.

Reproduction

Etcd3ClusterStorage.delete() is declared to return bool and the shared ClusterStorage contract says delete returns whether the key was deleted, but the etcd implementation discards the client result:

async def delete(self, key: str) -> bool:
    self.client.delete(key)

The successful path therefore always returns None, unlike the HTTP backend and unlike the method contract.

Expected behavior

Return the boolean result from the etcd client's delete() call.

Actual behavior

Successful etcd deletes return None.

Proposed fix

Return self.client.delete(key) and add a CPU-only mocked regression test covering both True and False client results.

Start at Etcd3ClusterStorage.delete() and compare it with the shared ClusterStorage contract and HTTP backend behavior. Add a CPU-only mocked regression test covering both True and False client results; done means the etcd delete returns the client's boolean result and the test passes.