Skip to main content
Team LeaderLuca Bigoni

Evolving toward cloud-native microservices

1 min read
microservicescloud-nativegokubernetes

The context

A monolithic backend that works isn't inherently a problem. It becomes one when the cost of every change grows faster than the value it delivers: increasingly risky deploys, teams stepping on each other's toes in the same codebase, scalability concerns that affect the whole system even when only one part of it needs to scale.

An evolution, not a rewrite

When those signals become constant, the approach I find most sustainable is an incremental evolution toward cloud-native microservices, for example on Kubernetes, rather than a big-bang rewrite. Every extracted service should have a clear domain boundary and its own data schema, and communicate with other services via gRPC for synchronous calls and events for asynchronous ones.

The trade-offs

  • More operational complexity, in exchange for independent, safer deploys.
  • More network latency between services, to be mitigated with well-defined gRPC contracts and end-to-end observability.
  • More explicit team coordination, with clearer ownership boundaries.

The main lesson

Observability and integration testing aren't an afterthought: they need to be designed alongside the architecture, not bolted on later. That investment is what makes it possible to evolve a distributed system with confidence.