Myself and Dragon discussed this a while back (specifically around retries for failed provisioning attempts and orchestration). I proposed the fanout solution, but Dragon talked me into the multiple notifiers scheme. I think it's a better solution, less to clean up if something fails.
Needs a proper test though. Specifically around one driver failing and the others continuing to work.
Myself and Dragon discussed this a while back (specifically around retries for failed provisioning attempts and orchestration). I proposed the fanout solution, but Dragon talked me into the multiple notifiers scheme. I think it's a better solution, less to clean up if something fails.
Needs a proper test though. Specifically around one driver failing and the others continuing to work.