>Microservices takes at least 25% of your engineering time just maintaining the platform. It's not worth it unless you can recoup that 25% in efficiency. My ta

>Microservices takes at least 25% of your engineering time just maintaining the platform. It's not worth it unless you can recoup that 25% in efficiency. My take is, microservices multiply your operational problems so you need a really solid platform/infrastructure/developer experience team otherwise you'll introduce a bunch of new headaches. These include things like CI/CD (building, storing artifacts, testing artifacts, promoting artifacts between environments), observability (metrics, logs, distributed tracing, error monitoring), distributed systems quirks like cascading failures, release/change management and communication, service discovery/routing, automated infra provisioning for things like datastores. Unless you have a really good handle on all these pieces (which I think most startups don't), you end up in an operational nightmare spending tons of time trying to keep the system going. Of course, the next step is throwing on 3rd party products to try to solve this which adds even more complexity. Throw in Datadog for observability, Istio service mesh for networking/traffic, ArgoCD/kustomize/Helm/Kubernetes to manage all this infra, etc etc