State Management

Overview

State management defines where application state lives, how it changes, and how consistency is preserved across distributed components.

State Categories

• Request state: short-lived per operation
• Session state: user context across requests
• Domain state: persistent business records
• Derived state: computed projections and caches

flowchart TD
    A[Client Request] --> B[API Service]
    B --> C[(Primary Store)]
    B --> D[(Cache Store)]
    C --> E[Event Stream]
    E --> F[(Read Model)]

Practical Example

A shopping application keeps cart session state in Redis and order state in PostgreSQL.

cartState:
  store: redis
  ttlSeconds: 1800
orderState:
  store: postgresql
  consistency: strong

This split keeps carts responsive while preserving transactional correctness for payments and orders.

Design Techniques

• Keep services stateless when possible
• Externalize session state for horizontal scaling
• Use optimistic concurrency for conflicting updates
• Emit domain events to synchronize projections

Common Pitfalls

• Storing critical business state only in cache
• Ignoring conflict handling in concurrent updates
• Mixing transient and durable state without clear ownership

Key Takeaways

• State location directly affects scalability and correctness.
• Session and domain state usually require different storage strategies.
• Clear ownership and consistency rules prevent data anomalies.