Improve retry wrapper behavior

[redreceipt]

Summary

• raise the final exception in with_retries so callers know a job failed
• remove the run_job scheduler wrapper so errors surface directly

Testing

• pip install -r requirements.txt
• pip install -r requirements.dev.txt (fails: Tunnel connection failed)
• flake8 (command not found)

---

https://chatgpt.com/codex/tasks/task_e_688a829406548324b51705ac7564e48d

[Copilot]

Pull Request Overview

This PR improves error handling in the retry wrapper by ensuring that the final exception is raised to the caller when all retry attempts are exhausted, rather than silently failing.

• Adds explicit exception raising after the final retry attempt fails
• Removes silent failure behavior that would prevent callers from knowing a job failed

[Copilot]

The hardcoded value '2' creates a magic number that's unclear. Consider using a variable like 'max_retries - 1' or adding a comment to explain that this represents the final retry attempt.