Microservices Are Over: Why We Moved Back To A Monolith And Cut P95 Latency By 80%

The request was not slow because our code was bad. It was slow because one simple user action had to survive too many network hops before… Continue reading on Medium »