You do have to wonder how long #Twitter can stay up, and it's not going to be gradual. Anyone that’s worked in ops knows that there is always a system that is never right, and needs to be manually unfucked every day. The fastest I’ve seen something fly apart when the only person that knew how it worked left the office to go on holiday was 12 minutes. #TwitterLayoffs#TwitterMigration
@aallan I think the site will just get progressively more unbearable over time due to random timeouts in malfunctioning subsystems. That's what I'd expect from an untended microservice arch.
@richunger It’s possible. That said, microservice architectures taking on technical debt tend toward big blobs of stuff, which tend to fail catastrophically. With so many people looking at bits of the code base they’ve never looked at before. How much technical debt do you think they can acquire in record time?
@aallan There's lots of ways for microservice architectures to take on debt. I've heard of several where it's the opposite. They have 3x as many services as engineers, and aren't investing in making service ownership mostly automated.
@richunger It’s possible. That said, microservice architectures taking on technical debt tend toward big blobs of stuff, which tend to fail catastrophically. With so many people looking at bits of the code base they’ve never looked at before. How much technical debt do you think they can acquire in record time?