backend-patterns

Use resource-based URLs like GET /api/markets/:id for single resources and query parameters for filtering, sorting, and pagination. Wrap handlers with middleware functions for authentication and centralize error handling in a shared errorHandler function.

Batch fetch related records instead of querying in a loop. Collect all foreign keys, run one query to fetch the related rows, then build a Map keyed by ID to join the results in memory.

The Repository pattern abstracts data access behind an interface like MarketRepository with methods such as findAll and findById. Concrete implementations like SupabaseMarketRepository handle the actual queries, keeping business logic independent of the database.

Yes, check Redis for a cached key first, and on a miss query the database and store the result with setex and a TTL such as 300 seconds. Invalidate the key with redis.del when the underlying record changes.

Map roles like admin, moderator, and user to permission arrays, then use a higher-order requirePermission function that verifies the JWT, checks the permission, and only then invokes the handler. Unauthorized requests return 403 responses.

Use an in-memory job queue for non-blocking tasks like indexing after a resource is created, so the API responds immediately. For distributed or persistent workloads, a simple in-process queue is not sufficient and a dedicated queue system is needed.