Add timeout sensor for long Dagster runs

[andrewm4894]

Summary

• stop zombie Dagster runs with new kill_long_running_runs sensor
• provide helper script kill_zombie_runs.py
• expose a kill-zombies command in the Makefile
• register the sensor in anomstack.main

Testing

• ruff check . (fails: Found 70 errors)
• make tests

---

https://chatgpt.com/codex/tasks/task_e_686998dd6f288328a9c67758380e26bf

[Copilot]

Pull Request Overview

Introduces automated and manual cancellation for long-running Dagster runs.

• Adds terminate_long_running_runs logic and a kill_long_running_runs sensor to auto-cancel overdue runs
• Provides a scripts/kill_zombie_runs.py helper and a kill-zombies Makefile target for manual invocation
• Registers the new sensor in anomstack.main to enable on startup

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
scripts/kill_zombie_runs.py	adds manual cancellation script invoking timeout logic
anomstack/sensors/timeout.py	implements terminate_long_running_runs and sensor logic
anomstack/main.py	registers the kill_long_running_runs sensor
Makefile	defines kill-zombies command for the helper script

Comments suppressed due to low confidence (2)

Makefile:88

• [nitpick] The Makefile target kill-zombies, script kill_zombie_runs.py, and sensor kill_long_running_runs use inconsistent naming conventions; consider unifying to a single style for clarity.

kill-zombies:

anomstack/sensors/timeout.py:50

• No tests were added for the terminate_long_running_runs function or the kill_long_running_runs sensor; consider adding unit tests to validate timeout and cancellation behavior.

@sensor(name="kill_long_running_runs", minimum_interval_seconds=CHECK_INTERVAL_SECONDS)

[Copilot]

[nitpick] Consider adding a shebang (e.g., #!/usr/bin/env python3) at the top of this script to make it directly executable.

[github-actions]

📊 Test Coverage Report

Coverage: 63% (yellow)

✅ Coverage maintained or improved!

💡 See detailed coverage report in the tests README