VM logs are hard to get in case of failures so it would be better if the logs are pushed to a redis instance so we can inspect the logs easier.
(redis because tailstream supports only redis and a ws. maybe we should support another sink in tailstream)