Few architecture questions cause more debate than this one. Microservices promise independent deployments and teams that move fast. Monoliths promise simplicity. Both promises are true, in the right context.
What microservices really cost
Each service needs its own pipeline, monitoring, data store and on-call ownership. Calls that were function calls become network calls that can fail or slow down. Data that lived in one database now has to stay consistent across several. These costs are worth paying when you have many teams; they are a heavy tax when you have one.
A simple decision guide
- One or two teams, product still finding its shape: start with a modular monolith
- Several teams blocking each other on releases: split out the most independent areas
- One component with very different scaling needs, such as video processing: run it as a separate service
- Strict isolation needs for compliance: separate the regulated part
The modular monolith
A modular monolith is one deployable application with strong internal boundaries: each module owns its data and exposes a clear interface to the others. You get the simplicity of one deployment and the option to extract a module into a service later, because the seams are already there.
Signs it is time to split
Look for evidence, not fashion: release queues where teams wait on each other, a single module causing most incidents, or scaling costs driven by one workload. When those appear, extract one service at a time and measure the result before moving on.
Our engineers and consultants write about what they see in real projects. Have a question about your own systems?






