Does anyone have experience adding redis oplog to an existing production deployment? What are the potentiall “gotchas” here. We’re hosting in AWS EC2 with self-hosted mongo too and would do the same thing with redis. We have a handfull of processes that modify the DB from outside of Meteor, I already know we’ll need to update these to notify redis of the change, but is there anything else?
What happens if a redis server goes down, do you just lose reactivity, or do you lose the ability to read too?