[morgante]

I think this level of expense is totally acceptable and justified when you consider how expensive developer time is today. Even for someone fluent in operations, I'd argue it doesn't make sense to cost optimize until you're spending $1,000+ monthly.

In the entire lifetime of their service, they've spent $8,727.57. That's less than it would cost to hire me to optimize their infrastructure.

Even though I've worked on building cost-optimized and efficient deployments for other companies, for my own projects I first turn straight to Heroku. It requires substantial traffic for my opportunity cost to be justified: I'm typically much better off charging larger companies to tune their deployments than using the same skills to shave pennies off my bill.