Design Food Ordering Flow - System Design | WinJob

Problem Statement: Consumer Food Ordering Flow

Design the consumer food ordering flow for DoorDash/Uber Eats/Grubhub: discovery, customization, pricing, and order placement with a frozen snapshot handed to logistics.

Lens	Detail
Primary actor	Hungry customer on mobile/web
Write path	Cart mutations → quote → checkout command
Read path	CDN menu + search index + cart cache
Consistency	AP browse, CP at payment boundary

Section focus (sec-001): three-sided marketplace with menu browse, modifiers, and checkout handoff to fulfillment. The core mechanism here is order command service. Track checkout_p95_ms in dashboards; page when dinner-peak checkout error rate exceeds 0.8% for 5 minutes.

Production scar: stale menu price at pay click. Mitigate with server-side validation, pinned menu_version_id, and idempotent checkout keys stored in Postgres with a unique index.

Operational narrative

Customer C-8001 adds a bowl with extra protein (+$2.50), applies promo LUNCH15, and tips 18%. Quote service recomputes tax using merchant nexus table T-NYC-04 before the client shows a pay button. If the kitchen marks the item 86'd between quote and pay, checkout returns 409 ITEM_UNAVAILABLE with a diff payload—not a silent price change.

Deep dive

At 387 checkout requests/sec peak, shard carts by user_id mod 256 and orders by restaurant_id + day bucket. Never let fulfillment APIs mutate cart lines; only the order command service writes orders rows after payment authorization succeeds.

javaOne Dark Pro

1 public record CheckoutCommand(String cartId, String menuVersionId, String idempotencyKey, Money authorizedTotal) {}

pythonOne Dark Pro

1 def validate_modifier_group(selection: list[str], rules: dict) -> None:
2     if len(selection) < rules['min'] or len(selection) > rules['max']:
3         raise ValueError('modifier constraint violated')

typescriptOne Dark Pro

1 export interface CartLine { skuId: string; quantity: number; modifiers: string[]; unitPriceCents: number; }

Interview checkpoints (sec-001)

Why interviewers care

Food Ordering Flow interviews reward crisp scope, explicit trade-offs, and failure stories—not generic microservice diagrams.

Interview checkpoint

Name one failure story for Problem Statement: Consumer Food Ordering Flow that proves you understand real outages, not happy-path diagrams.

Key Highlights

•order command service anchors Problem Statement: Consumer Food Ordering Flow
•Metric: checkout_p95_ms
•Guard: stale menu price at pay click
•understanding phase checkpoint

1	def validate_modifier_group(selection: list[str], rules: dict) -> None:
2	if len(selection) < rules['min'] or len(selection) > rules['max']:
3	raise ValueError('modifier constraint violated')