Floating IP and BW on HA droplets

I have a question on bandwidth.

Lets say I have the requirement for 2x application servers which are $160/6TB transfer.

Now I want to take advantage of the floating IP and I get 2x $5 boxes because HA Proxy doesn’t need much resources. The $5 is 1TB transfer.

Can anyone see what’s wrong here? I’m only getting 1TB transfer so these boxes would be done pretty soon and 5TB would be left on the app servers.

Would I want to get all the same servers for this type of setup? No way, I’d have 2 servers effectlvely sitting idle and costing $$$.

So what would the solution here be? I would rather have some sort of load balancer virtual device and one that would scale with increase of traffic. Something AWS still doesn’t do!

Would love to hear from DO on what they recommend! I’m looking to put this into prod in Jan. AWS, DO and Vultr are competing for my business here. Used AWS before didn’t like, heard great things about DO.