> That’s why these systems are built the way they are. Built how? Because I can state with confidence that I have cleaned up a ton of failed upgrades/zombie te

> That’s why these systems are built the way they are. Built how? Because I can state with confidence that I have cleaned up a ton of failed upgrades/zombie terraform deploys of these serverless kubernetes wonders that followed every best practice under the sun, and these things are not really considered even moderately reliable (as designed by imperfect mortals under real world conditions), meanwhile professionally, people who stand to lose a lot of money should their systems go down generally operate systems whose architectures were designed decades ago, are generally horizontally scaled monoliths, and are extremely conservative in software choice. Also downtime often is no biggie, as long as it's planned and or don't lose (too much) customer critical data. Like nodobody cares if your test db cluster goes down for the weekend. We even shut down our db instances to save money.