Skip to content

Handle API Status errors 499 differently #132

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
p- wants to merge 7 commits into from
Closed

Handle API Status errors 499 differently #132

Changes from 3 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
47ae16d
Retry on 499 errors
p- Jan 20, 2026
70c8be7
add logging statement
p- Jan 20, 2026
c38b939
Add APIStatusError handler
p- Jan 20, 2026
a0fe49e
Improve error logging for API status errors
p- Jan 20, 2026
d665364
Merge branch 'main' into p--handle-499
kevinbackhouse Feb 5, 2026
389529a
Merge branch 'main' into p--handle-499
kevinbackhouse Feb 9, 2026
14f2480
Merge branch 'main' into p--handle-499
kevinbackhouse Feb 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 16 additions & 2 deletions src/seclab_taskflow_agent/__main__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,7 @@
from agents.mcp import MCPServer, MCPServerStdio, MCPServerSse, MCPServerStreamableHttp, create_static_tool_filter
from agents.extensions.handoff_prompt import prompt_with_handoff_instructions
from agents import Tool, RunContextWrapper, TContext, Agent
from openai import BadRequestError, APITimeoutError, RateLimitError
from openai import BadRequestError, APITimeoutError, RateLimitError, APIStatusError
from openai.types.responses import ResponseTextDeltaEvent
from typing import Callable

Expand Down Expand Up @@ -334,7 +334,16 @@ async def _run_streamed():
return
except APITimeoutError:
if not max_retry:
logging.error(f"Max retries for APITimeoutError reached")
logging.error(f"Max API retries reached")
raise
max_retry -= 1
except APIStatusError as e:
# Retry transient “client closed request / upstream cancelled” style errors
if getattr(e, "status_code", None) != 499:
raise # propagate non-499 errors
# 499: retry
if not max_retry:
logging.error(f"Max API retries reached")
raise
max_retry -= 1
Comment on lines +358 to 366
Copy link

Copilot AI Jan 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The new APIStatusError exception handler should be added to the outer exception handling block (lines 361-381) to provide consistent error handling and user feedback. Currently, if a non-499 APIStatusError is raised and propagated from the inner try-except block, it won't be caught by any outer handler, potentially resulting in an unhandled exception. Consider adding an exception handler for APIStatusError similar to the existing handlers for BadRequestError and APITimeoutError.

Copilot uses AI. Check for mistakes.
Copy link
Copy Markdown
Contributor

@m-y-mo m-y-mo Jan 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yea probably worth adding a handler in the outer scope like APITimeoutError

Comment on lines +358 to 366
Copy link

Copilot AI Jan 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The new APIStatusError retry logic for 499 status codes lacks test coverage. Since the repository includes tests for other functionality (test_api_endpoint_config.py, test_cli_parser.py, test_yaml_parser.py), consider adding tests to verify that 499 errors are properly retried and that non-499 APIStatusErrors are propagated correctly.

Copilot uses AI. Check for mistakes.
Comment on lines +358 to 366
Copy link

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In the APIStatusError(499) retry path, the max-retries case logs with logging.error and re-raises without preserving traceback/context. Consider logging with traceback (or include exception details) before raising so repeated 499 failures are diagnosable.

This issue also appears on line 353 of the same file.

Copilot uses AI. Check for mistakes.
except RateLimitError:
Expand Down Expand Up @@ -370,6 +379,11 @@ async def _run_streamed():
async_task=async_task,
task_id=task_id)
logging.error(f"Bad Request: {e}")
Copy link

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The APITimeoutError outer handler logs Bad Request: {e} even though this is a timeout path. This is misleading when debugging and in logs/telemetry; update the log message to reflect a timeout (and consider logging.exception if you want traceback).

Suggested change
logging.error(f"Bad Request: {e}")
logging.exception("Timeout Error")

Copilot uses AI. Check for mistakes.
except APIStatusError as e:
await render_model_output(f"** 🤖❗ API Status Error: {e}\n",
async_task=async_task,
task_id=task_id)
logging.error(f"Bad Request: {e}")

if async_task:
await flush_async_output(task_id)
Expand Down