I thought about that but ultimately decided that obscuring it didn’t really make sense as it is a noninvasive tracker, but depending on the business, this could make sense.
Yea, an HA Proxy would make a lot of sense with something like this. As you scale up the number of Interference services for SD, you can increase the number of background workers that are available on the Ruby side to take advantage of the additional servers. You can probably set up HA Proxy to balance the load which would help simplify the infrastructure. By default, I think it does Round Robin which would route each task to a different server in turns.
The logic that I meant was for sending out the emails. Typically, instead of looping over all of the users, you would check if the user has unsubscribed from the email, if there was any kind of time preference, using batch finds instead of looping through every user, etc.
