Error handling issues in `EventBridgeSchedulerHelper`

[Mattral]

Error handling issues in EventBridgeSchedulerHelper

Summary

There are a few small but impactful error-handling issues in
src/sagemaker/mlops/workflow/_event_bridge_client_helper.py that can lead to unexpected exceptions or reduced debuggability.
These can be fixed safely without changing public behavior or impacting other parts of the codebase.

---

Affected File

sagemaker-mlops/src/sagemaker/mlops/workflow/_event_bridge_client_helper.py

---

Issues Identified

1. delete_schedule does not handle non-existent schedules

The delete_schedule method unconditionally calls delete_schedule on the EventBridge Scheduler client.

If the schedule does not exist, AWS raises a ResourceNotFoundException, even though the docstring states:

“Deletes an EventBridge Schedule of a given pipeline if there is one.”

This results in an unexpected exception for callers.

Expected behavior
Deleting a non-existent schedule should be a no-op or gracefully handled.

---

2. Inconsistent and unused error-code constants

The file defines two constants:

RESOURCE_NOT_FOUND = "ResourceNotFound"
RESOURCE_NOT_FOUND_EXCEPTION = "ResourceNotFoundException"

Only RESOURCE_NOT_FOUND_EXCEPTION is used.
RESOURCE_NOT_FOUND is unused and can be safely removed to avoid confusion and improve code clarity.

---

3. Exception re-raise style loses traceback context

In upsert_schedule, the exception is re-raised using:

raise e

This resets the traceback and makes debugging more difficult.

---

Why This Matters

• Prevents unexpected runtime failures when deleting schedules
• Improves debuggability by preserving tracebacks
• Cleans up unused constants for better maintainability
• All fixes are low-risk and backward-compatible

---

Suggested Fix Scope

• Catch and ignore ResourceNotFoundException in delete_schedule
• Remove unused RESOURCE_NOT_FOUND constant
• Replace raise e with raise to preserve traceback

---