Internally here at Aptible, 4-5 TBs in size...Our customers however? The biggest is > 14 TB. Now when we get into the RAM side of things? More than AWS supports via RDS both internally and in our Customer's Environments. What's fascinating is we see people scaling db's up endlessly without understanding why, which is typically a red flag of sorts. Then we reach out and find out it's just an unused replication slot and Boom, save them a bunch of money in an instant. Plus TONS of money long term since they modeled all their cost projections assuming it was valid usage. :)