backend-patterns

Use resource-based URLs like GET /api/markets for lists and GET /api/markets/:id for single resources, with query parameters for filtering, sorting, and pagination. Separate concerns into repository, service, and controller layers for maintainability.

Batch fetch related records instead of querying in a loop. Collect all foreign keys, fetch related records in one query, then map them back in memory using a Map keyed by ID.

The repository pattern abstracts data access behind an interface, so business logic depends on the interface rather than a specific database client. This lets you swap implementations, such as a Supabase repository for a cached one, without changing service code.

Yes, cache-aside works by checking Redis first, fetching from the database on a miss, then populating the cache with a TTL. The pattern shown uses setex with a 300-second expiry and explicit invalidation on updates.

Use middleware like a withAuth wrapper for cross-cutting concerns applied to many routes, such as token verification. Use the service layer for business-rule authorization, like role-based permission checks tied to specific operations.

Exponential backoff spaces retries at increasing intervals (1s, 2s, 4s) to avoid overwhelming a recovering service. It handles transient failures gracefully while still surfacing persistent errors after the maximum retry count.