HTTP & REST basics for model inference APIs — Questions & Answers

Test your understanding of HTTP and REST fundamentals as they apply to model inference APIs, including status codes, idempotency, pagination, and request validation. This quiz is ideal for anyone looking to solidify their grasp of core concepts for reliable and efficient API usage.

This quiz contains 15 questions. Below is a complete reference of all questions, answer choices, and correct answers. You can use this section to review after taking the interactive quiz above.

1. Question 1: Understanding Successful Requests

   Which HTTP status code indicates that a model inference request was successful and the result is included in the response body?

   • 500 Internal Server Error
   • 404 Not Found
   • 301 Moved Permanently
   • 200 OK
   Show correct answer

   Correct answer: 200 OK

2. Question 2: Identifying Missing Resources

   If you send a GET request to an inference API for a resource that does not exist, which status code is most appropriate for the response?

   • 404 Not Found
   • 403 Forbidden
   • 201 Created
   • 202 Accepted
   Show correct answer

   Correct answer: 404 Not Found

3. Question 3: Creating Inference Tasks

   When a new inference job is successfully created using a POST request, which status code should the API return?

   • 304 Not Modified
   • 400 Bad Request
   • 201 Created
   • 200 OK
   Show correct answer

   Correct answer: 201 Created

4. Question 4: Recognizing Bad Requests

   Which HTTP status code is returned if the input data for a model inference API is missing a required field?

   • 400 Bad Request
   • 204 No Content
   • 410 Gone
   • 307 Temporary Redirect
   Show correct answer

   Correct answer: 400 Bad Request

5. Question 5: Understanding Idempotency

   Which HTTP method is considered idempotent and will yield the same result if used multiple times for a model prediction resource?

   • CONNECT
   • TRACE
   • POST
   • GET
   Show correct answer

   Correct answer: GET

6. Question 6: POST and Idempotency

   When calling a model inference API, why is a POST request generally not idempotent?

   • POST always deletes existing data
   • Each POST request can create a new resource or trigger a new operation
   • POST is restricted to retrieving information
   • POST cannot return a response body
   Show correct answer

   Correct answer: Each POST request can create a new resource or trigger a new operation

7. Question 7: Understanding Pagination

   Why might a model inference API implement pagination in its GET responses when listing predictions?

   • To prevent overwhelming clients with too much data at once
   • To randomize the order of responses
   • To avoid supporting different status codes
   • To reduce the number of HTTP methods
   Show correct answer

   Correct answer: To prevent overwhelming clients with too much data at once

8. Question 8: Pagination Parameters

   Which parameter is commonly used in REST APIs to indicate the maximum number of results per page when retrieving model inference outputs?

   • locate
   • retry
   • delete
   • limit
   Show correct answer

   Correct answer: limit

9. Question 9: Specifying the Next Page

   After receiving a paginated response from an inference API, which parameter would you use to request the next set of results?

   • offset
   • close
   • clear
   • remove
   Show correct answer

   Correct answer: offset

10. Question 10: Status Code for Unauthorized Access

    If a user tries to access a prediction resource without proper authentication, which HTTP status code should the API respond with?

    • 418 I'm a teapot
    • 409 Conflict
    • 202 Accepted
    • 401 Unauthorized
    Show correct answer

    Correct answer: 401 Unauthorized

11. Question 11: Proper Request Validation

    What should a REST API do when it receives an inference request with a valid structure but invalid data type for an input field?

    • Return a 201 Created status
    • Ignore the invalid field and proceed
    • Return a 400 Bad Request error indicating validation failure
    • Return a 200 OK with an empty response
    Show correct answer

    Correct answer: Return a 400 Bad Request error indicating validation failure

12. Question 12: Safe HTTP Methods

    Which HTTP method is considered 'safe' because it does not modify resources, as when retrieving prediction logs?

    • DELETE
    • GET
    • PUT
    • PATCH
    Show correct answer

    Correct answer: GET

13. Question 13: Reporting Server Errors

    If the inference model server encounters an unexpected condition preventing it from fulfilling a request, which status code should it return?

    • 405 Method Not Allowed
    • 500 Internal Server Error
    • 301 Moved Permanently
    • 204 No Content
    Show correct answer

    Correct answer: 500 Internal Server Error

14. Question 14: Consistent Request Results

    Which statement best describes an idempotent API operation for model inference?

    • Any request must create new resources each time
    • Performing the same operation multiple times results in the same effect
    • Each request must return a unique identifier
    • The server always responds with a 404 status code
    Show correct answer

    Correct answer: Performing the same operation multiple times results in the same effect

15. Question 15: Method for Updating Prediction Settings

    To update the configuration of a model inference task without creating a new one, which HTTP method should you use?

    • OPTIONS
    • PUT
    • GET
    • POST
    Show correct answer

    Correct answer: PUT