RPC Error Handling and Status Codes
You'll learn how RPC systems communicate errors to clients using structured responses and standardized status codes.
What Is RPC Error Handling?
When you call a remote procedure, things can go wrong: the server might be down, the procedure might not exist, or your arguments might be invalid. Unlike REST APIs where you rely on HTTP status codes, RPC systems need their own error-handling mechanisms since they can run over different transport protocols (not just HTTP).
RPC frameworks use structured error responses that typically include:
- A status code (similar to HTTP's 404 or 500)
- A message describing what went wrong
- Optional details like stack traces or error metadata
Framework-Specific Approaches
JSON-RPC uses standardized error codes in its response object:
- Code
-32700: Parse error (invalid JSON) - Code
-32600: Invalid request - Code
-32601: Method not found - Custom codes for application-specific errors
gRPC (which we introduced in lesson 799) defines its own status codes:
OK(0): SuccessINVALID_ARGUMENT(3): Client sent bad dataNOT_FOUND(5): Resource doesn't existINTERNAL(13): Server-side error
These codes travel in the RPC protocol's metadata, separate from the actual response payload.
Why This Matters
Think of status codes as a universal language for "what went wrong." Just like traffic lights use red/yellow/green everywhere, RPC status codes let clients understand errors without parsing cryptic messages. A well-designed error response tells the client whether to retry, fix their request, or alert a human.
Key Takeaway: RPC error handling uses framework-specific status codes and structured error objects to communicate failures clearly and consistently, enabling clients to respond appropriately to different error conditions.