pytorch/serve

Custom Exceptions Are Ignored By Workflow

Open

#2,456 opened on Jul 11, 2023

View on GitHub
 (0 comments) (3 reactions) (0 assignees)Java (3,844 stars) (790 forks)batch import
help wantedtriagedworkflowx

Description

I have a workflow and defined custom exceptions in model handlers for error handling. For an example exception, I have got desired error response when I send request directly to the model in the workflow, using API predictions/workflowName__modelName. The problem is when sending to workflow, exception code and message replaced with fixed code, which is 500, and message: "Error executing ModelName". How can I configure workflow to make it return the same exception that the model raises.

What I implemented in model handler: "download".

try:
    _, image_file = get_image_from_url(url)
except Exception as e:
    raise PredictionException(f'Error downloading image. Error: {e}.', 501)

What I got from requesting directly to the model (predictions/workflowName__download). This is what I want to see when I requested to workflow.

{
    "code": 501,
    "type": "InternalServerException",
    "message": "Error downloading image. Error: Not Found Image URL https://example_url"
}

What workflow returns.

{
    "code": 500,
    "type": "InternalServerException",
    "message": "Error executing download"
}

Contributor guide

Tech stack
pythonjava
Domain
backendapi
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
Understanding of PyTorch Serve model handlersFamiliarity with workflow execution
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
55
Research direction
Investigate the workflow executor code in PyTorch Serve, specifically how exceptions from model handlers are caught and transformed. Look at the relevant source files (e.g., workflow.py or similar) to understand where the custom exception code and message are being replaced with generic 500 errors. Consider if there is a configuration option to propagate exceptions or if a fix is needed in the exception handling logic.