Integrations, Data Flow, and Extensibility
What to know
Understand integration patterns: REST and GraphQL APIs, SOAP (legacy), webhooks, message queues (async), and scheduled imports/exports. Know authentication (integration tokens, OAuth for admin, customer tokens for GraphQL). Understand service contracts as the API surface and how repositories map to endpoints. Know message queue concepts for async order/inventory sync and cron for scheduled jobs.
Data flow: product/customer/order data moving between Commerce, ERP, PIM, CRM, and payment gateways. Understand idempotency, retry, and error handling; bulk vs. real-time sync; and middleware roles. Extensibility: plugins, observers, custom modules, declarative schema, and GraphQL schema extension for headless storefronts. Know headless architecture (PWA Studio, GraphQL) vs. traditional storefront.
Traps
Questions confuse REST vs. GraphQL use cases (GraphQL for storefront/headless efficiency; REST for admin/integration CRUD). Another trap: assuming synchronous real-time sync is reliable at scale - bulk/async with queues is preferred. Watch token scopes and that integration tokens differ from customer tokens. Also beware assuming all data is exposed via API without service contract coverage.
Rule of thumb
For high-volume, failure-tolerant data movement, prefer asynchronous message queues with idempotent consumers; for interactive storefront queries, prefer GraphQL.
Worked example
An ERP pushes 50,000 price updates nightly. Use a scheduled bulk import (async) rather than 50,000 real-time REST calls, then reindex.