frameworkEngineering & architecture

Microservices Architecture

An architectural style that divides a system into independently deployable services aligned to cohesive business capabilities and owned end to end.

Download Markdown (.md)

01What it means

Microservices architecture structures a product as a set of small, autonomous services communicating through networked interfaces. Each service encapsulates a cohesive business capability, can usually be deployed independently, and owns the behavior and data needed for its responsibilities. “Micro” describes focused scope and autonomy, not a required line count.[1][2][3]

Autonomous business capabilities Independent services collaborate through explicit contracts
APIs and events
Service A
Capability + owned data
Service B
Capability + owned data
Service C
Capability + owned data
Delivery automation
Independent deployment
Platform capabilities
Discovery, telemetry, secrets
Ownership
Operation and lifecycle

Focused scope and operational autonomy matter more than service count.

02Why it matters

Independent services can let teams release, scale, secure, and evolve different capabilities according to their needs. Faults and changes may have a smaller blast radius, and ownership can align with business capabilities. These benefits become meaningful mainly when organizational structure, delivery automation, observability, and service boundaries support real independence.

03When to use it

Consider microservices for a sufficiently complex, growing system where multiple teams need independent delivery or where capabilities have sharply different scaling and reliability needs. Start with evidence of those pressures. A monolith is often safer for a small team, an uncertain product, or a domain whose boundaries are not understood.

04How it works

  1. Identify cohesive business capabilities and boundaries through domain analysis rather than splitting by technical layer.
  2. Assign clear end-to-end ownership for each service, including operation, data, security, and lifecycle.
  3. Define explicit APIs or events and design them for compatibility, failure, latency, and independent change.
  4. Automate build, test, deployment, infrastructure, policy checks, and rollback or recovery for every service.
  5. Provide shared platform capabilities for discovery, telemetry, secrets, networking, and incident response.
  6. Continuously reassess service boundaries and combine or split services when operational evidence justifies it.

05A practical example

Example

A travel platform separates availability search from booking because search has extreme read traffic and can tolerate slightly stale data, while booking requires strict consistency and auditability. Separate teams deploy and scale them independently through a versioned interface, but keep tightly coupled booking rules together within one service.

06Where it breaks

Microservices trade in-process simplicity for distributed-system complexity: latency, partial failure, data consistency, security surfaces, testing, and operational overhead. Poor boundaries create chatty calls and coordinated releases—the costs of distribution without autonomy. More services also mean more pipelines, telemetry, dependencies, and cognitive load for teams to manage.

Common mistakes

  • Splitting a young product into services before its domain and team boundaries are understood.
  • Creating one service per database table, endpoint, or small function instead of a cohesive capability.
  • Sharing databases or requiring synchronized releases while claiming independent ownership.
  • Adopting microservices without investing in automated delivery, observability, and operational support.

07Sources

  1. MicroservicesMartin Fowler · Foundational practitioner articleMicroservices
  2. Microservices GuideMartin Fowler · Expert reference collectionMicroservices Guide
  3. Use Domain Analysis to Model MicroservicesMicrosoft Learn · Institutional architecture guideUse Domain Analysis to Model Microservices
View saved topics

Start typing to search titles, definitions, categories, and frameworks.