Error handling and rate limits in Prompt Engineering / GenAI - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the purpose of error handling in AI model APIs?

Error handling helps catch and manage problems like invalid inputs or server issues, so the system can respond gracefully without crashing.

beginner

Define rate limits in the context of AI services.

Rate limits control how many requests a user or app can send to an AI service in a set time, preventing overload and ensuring fair use.

beginner

Name two common types of errors you might handle when calling an AI API.

1. Client errors (like bad input data). 2. Server errors (like service downtime).

beginner

Why is it important to respect rate limits when using AI APIs?

Respecting rate limits avoids being blocked or slowed down, and helps keep the service stable for everyone.

intermediate

What is a common strategy to handle hitting a rate limit?

A common strategy is to wait (backoff) and retry the request after some time, often increasing wait time if repeated limits occur.

What does a 429 HTTP status code usually mean when calling an AI API?

A Request timeout

B Invalid API key

C Server internal error

D Too many requests - rate limit exceeded

A 429 status code means the client sent too many requests in a short time and hit the rate limit.

Which of these is NOT a good practice for error handling with AI APIs?

A Logging errors for debugging

B Retrying immediately without delay after an error

C Checking error codes to decide next steps

D Gracefully informing users about issues

Retrying immediately can cause repeated failures; it's better to wait before retrying.

What is the main goal of implementing rate limits on AI services?

A To make APIs harder to use

B To increase server costs

C To prevent abuse and ensure fair access

D To reduce the number of users

Rate limits help prevent abuse and keep the service stable and fair for all users.

If an AI API returns an error due to invalid input, what should your program do?

A Fix or ask for correct input before retrying

B Ignore the error and continue

C Retry the same request endlessly

D Shut down immediately

Fixing or requesting correct input helps avoid repeated errors and improves user experience.

What technique helps reduce hitting rate limits when making many AI API calls?

A Spreading requests over time (throttling)

B Sending all requests at once

C Using multiple invalid API keys

D Ignoring rate limit warnings

Throttling spreads requests to avoid bursts that exceed rate limits.

Explain why error handling and rate limits are important when working with AI APIs.

Think about what happens if too many requests come in or if something goes wrong.

Describe a simple approach to handle hitting a rate limit during AI API calls.

Consider how to be polite to the service and avoid repeated errors.