Implement a Resilient Aggregator over Three Services

Problem Statement

When a customer navigates to the checkout screen, the backend must construct a composite CheckoutContext by aggregating data from three separate downstream microservices:

1. UserService: get_user(user_id: str) -> UserProfile
   • Returns UserProfile(user_id: str, default_address_id: str, payment_customer_id: str).
   • Critical dependency: The other two services depend on IDs returned by this call.
2. AddressService: get_address(address_id: str) -> Address
   • Returns Address(address_id: str, formatted_address: str).
   • Non-critical: If this service fails or times out, the context should return address = None and record an error message without failing the entire checkout view.
3. PaymentService: get_payment_methods(customer_id: str) -> list[PaymentMethod]
   • Returns a list of PaymentMethod(method_id: str, card_type: str, last_four: str).
   • Semi-critical: Should retry once on network error (ClientNetworkError) before failing. If it still fails, the aggregator returns payment_methods = [] and records an error.

Implement the class CheckoutAggregator and its primary method:

class CheckoutAggregator:
    def __init__(
        self,
        user_service: UserClient,
        address_service: AddressClient,
        payment_service: PaymentClient,
    ):
        ...

    async def get_checkout_context(
        self, user_id: str, timeout_seconds: float = 1.0
    ) -> CheckoutContext:
        ...

Return Type

@dataclass
class CheckoutContext:
    user: UserProfile
    address: Optional[Address]
    payment_methods: list[PaymentMethod]
    errors: list[str]  # Non-fatal errors encountered during partial failure

Requirements

1. Dependency Execution: AddressService and PaymentService must run concurrently as soon as UserService finishes.
2. Critical Failure: If UserService raises an exception or times out, propagate a UserFetchError immediately and do not invoke downstream services.
3. Deadlines & Cancellation: If the overall execution exceeds timeout_seconds, any running downstream tasks must be promptly cancelled, and a TimeoutError should be raised if UserService was not resolved, or partial results returned if acceptable according to policy.
4. Retries: Retry PaymentService at most once on transient network errors before falling back to empty results.

Example

# User found, address succeeds, payment fails transiently then succeeds
context = await aggregator.get_checkout_context(user_id="usr_123", timeout_seconds=0.5)
assert context.user.user_id == "usr_123"
assert context.address.formatted_address == "900 Market St, SF"
assert len(context.payment_methods) == 1
assert context.errors == []