How do getResponse() and getStatus() work on an HttpException and what can getResponse() return?
getStatus() always returns a number — the HTTP status code. getResponse() returns either a string or an object depending on how the exception was constructed. When thrown with a string message it returns that string. When thrown with an object it returns that object. When thrown with no arguments it returns a default object with statusCode, message, and error fields.
- 1
Always guard with typeof res === 'string' before accessing .message — both shapes are valid.
- 2
getStatus() is always a number — safe to use directly without type checking.
- 3
Omitting the typeof check is the most common bug in custom exception filters.
- 4
When the body is an object, access nested fields like (body as any).message or use optional chaining.
- 5
exception.message is always a string — a safe alternative to getResponse() for simple use cases.
Scenario Questions
0-2 years experience
- 1If you catch a NestJS HttpException in a controller and call getStatus() and getResponse(), what values would you expect for a 404 Not Found error thrown with a custom message?
- 2Suppose you have a service that throws new HttpException('Invalid data', HttpStatus.BAD_REQUEST). In a unit test, how would you assert that getResponse() returns the string you passed versus an object?
2-5 years experience
- 1During a recent feature, you noticed that some HttpExceptions are returning an object with 'statusCode' and 'message' fields, while others return just a string. What could cause this inconsistency, and how would you standardize the response shape across the app?
- 2Your API gateway logs the status from getStatus() but sometimes the body from getResponse() is undefined. Walk me through how you would debug why getResponse() is empty for certain exceptions.
- 3You need to map different HttpException subclasses to specific error DTOs before sending them to the client. How would you use getResponse() to decide which DTO to instantiate, and what edge cases would you guard against?
5-8 years experience
- 1Design a global exception filter that logs the status and response of every HttpException, but also masks sensitive fields that might appear in getResponse(). What strategy would you use to detect and strip those fields without breaking the original payload?
- 2In a high‑traffic microservice, you want to avoid the overhead of serializing large error objects returned by getResponse(). How would you modify the exception handling pipeline to limit the size of the response while preserving getStatus() semantics?
- 3Explain how you would integrate NestJS’s HttpException handling with an external monitoring system that expects a flat error record. How does getResponse() shape affect the transformation, and what fallback would you implement for unexpected response types?
8+ years experience
- 1Across several teams, different modules throw HttpExceptions with varying response formats (string, object, array). Propose an organization‑wide contract and migration plan to unify getResponse() outputs, considering backward compatibility and CI enforcement.
- 2Your company is moving from NestJS to a new framework but wants to keep existing error handling semantics. How would you abstract getStatus() and getResponse() behavior so that legacy code can be ported with minimal changes, and what risks do you anticipate?
Follow-up Questions
- Can you give an example of a custom error object you’ve returned from getResponse()?
- What would happen if you called getStatus() on a non‑HttpException error?
- How does NestJS decide whether getResponse() returns a string versus an object?