Database per Service

Architectural Patterns

Meaning

A Database per Service pattern assigns each microservice its own dedicated database. It isolates data ownership, preventing cross‑service data coupling and reducing the blast radius of schema changes. It is used when services need independent scaling, deployment, or strict data encapsulation.

Primary Function

Database architecture

Communicative Purpose

Ensures each microservice manages its own data store, avoiding shared‑database contention and accidental data coupling.

Pattern

service → dedicated database → isolated data management

Core Structure

service_id → database_instance

Função primária

Database architecture

Propósito comunicativo

Ensures each microservice manages its own data store, avoiding shared‑database contention and accidental data coupling.

Situações de gatilho

Microservices: a new service requires its own schema and independent scaling; Domain‑driven design: bounded contexts need isolated persistence

Contextos

Microservice architectures, cloud‑native applications, containerized deployments, service‑oriented backends

Padrão

service → dedicated database → isolated data management

Estrutura central

service_id → database_instance

Colocados típicos

service
database
schema
connection pool
migration

Substituições comuns

Shared database with schema per tenant (simpler ops but tighter coupling)
multi‑tenant single database (resource efficient but riskier isolation)

Erros comuns

Creating a separate DB for every tiny service → operational overhead; Forgetting to configure migrations per service → schema drift; Using the same connection string across services → accidental data sharing

Similar / contraste

Shared Database (multiple services share one DB) – higher coupling; Multi‑tenant Database (single DB with tenant discriminator) – shared resources but logical isolation

Interferências

Coming from monolithic Java: assuming a single DataSource can be reused across services – leads to transaction leakage in microservices

Família do chunk

Database per Service
Database per Tenant
Shared Database
Multi‑tenant Architecture

Nuance

Do not use when the team is small and operational cost outweighs benefits; Extra databases increase connection and resource usage, impacting performance; Schema migrations must be coordinated per service to avoid version mismatches

Efeito pragmático

Reduces risk of cascading failures due to schema changes and enables independent deployment pipelines

Dica de memória

Think of each microservice as a house with its own backyard garden; the garden (database) belongs only to that house.

Nota

Database per Service often pairs with container orchestration tools that provision databases dynamically (e.g., Kubernetes Operators).

Upgrade path

Shared Database with schema segregation for low‑traffic services

Frequência: MediumFormulaicidade: Semi-fixedTipo de construção: conceptPrioridade de aquisição: Active recallPrioridade de output: BothTag de espaçamento: Medium-term

Examples

from flask import Flask, g import sqlite3 app = Flask(__name__) def get_db(): if 'db' not in g: g.db = sqlite3.connect('orders_service.db') return g.db @app.route('/orders') def list_orders(): cur = get_db().cursor() cur.execute('SELECT id, total FROM orders') return {'orders': cur.fetchall()} if __name__ == '__main__': app.run()

from fastapi import FastAPI, Depends from sqlalchemy import create_engine, Column, Integer, Float from sqlalchemy.orm import sessionmaker, declarative_base app = FastAPI() engine = create_engine('postgresql://user:pwd@localhost/payments_db') SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine) Base = declarative_base() class Payment(Base): __tablename__ = 'payments' id = Column(Integer, primary_key=True, index=True) amount = Column(Float) def get_db(): db = SessionLocal() try: yield db finally: db.close() @app.get('/payments') def read_payments(db: SessionLocal = Depends(get_db)): return db.query(Payment).all()

# settings.py (Django project with separate DB per service) DATABASES = { 'default': { 'ENGINE': 'django.db.backends.postgresql', 'NAME': 'inventory_db', 'USER': 'inventory_user', 'PASSWORD': 'pwd', 'HOST': 'localhost', 'PORT': '5432', }, 'shipping_db': { 'ENGINE': 'django.db.backends.postgresql', 'NAME': 'shipping_db', 'USER': 'shipping_user', 'PASSWORD': 'pwd', 'HOST': 'localhost', 'PORT': '5432', }, } # apps/shipping/models.py from django.db import models class Shipment(models.Model): tracking_code = models.CharField(max_length=100) class Meta: app_label = 'shipping' db_table = 'shipment' # routers.py from django.db import connections from .models import Shipment def list_shipments(): with connections['shipping_db'].cursor() as cursor: cursor.execute('SELECT id, tracking_code FROM shipment') return cursor.fetchall()