A software development team is integrating an external payment processing API that has strict limits on request rates and a defined payload size. The API is crucial for handling real-time financial…

Question

Options

• AIncrease the API timeout settings to accommodate higher latency during peak usage periods.
• BOpt for batch processing of transactions to reduce the number of API calls made.
• CImplement an adaptive load balancing mechanism that dynamically routes requests based on current API usage.
• DQueue transaction requests during peak times and process them at a controlled rate that complies with the API rate limits.

• A
  4% (1)
• B
  13% (3)
• C
  78% (18)
• D
  4% (1)

Explanation

There's actually a problem with this question: D is the stronger correct answer, not C.

Why D is correct: Queuing transactions and processing them at a controlled rate is the textbook solution for API rate limiting. It directly solves both stated problems - handling traffic spikes without violating rate limits or causing system failures. This pattern (often called a "token bucket" or "leaky bucket" queue) is standard practice in financial systems precisely because it absorbs bursts while guaranteeing compliance.

Why C is misleading: Adaptive load balancing routes requests across multiple internal servers, but a single third-party payment API with a global rate limit doesn't care which of your servers makes the call - the limit still applies. C would help if you had multiple API keys or endpoints, but the question doesn't state that, making this a questionable assumption.

Why the other options fail:

• A (increase timeout): Timeouts control how long you wait for a response, not how many requests you send - irrelevant to rate limits.
• B (batch processing): Conflicts with the "real-time" requirement stated in the question. Batching introduces latency that financial transactions can't tolerate.

Memory tip: When you see "rate limit + traffic spike," the answer is almost always a queue with rate control (throttling). If an answer mentions "load balancing" without mentioning multiple API endpoints/keys, treat it skeptically - you can't load-balance around a third-party quota.

If this is from a real exam, consider flagging the question, as D aligns better with industry practice and the stated constraints.

No community discussion yet for this question.