[API] endpoint to stop/kill a running job

[eshaan7]

In this case it would be nice to have a button - kill job or stop job.

Originally posted by @WhiteHatMR in https://github.com/intelowlproject/IntelOwl/issue_comments/707807962

We could add a kill button to the job (analysis) result page. This button when invoked should stop all ongoing celery tasks (i.e. all running analyzers) for the particular job. The status of such job should then be marked failed or killed.

This sounds easy on paper, but would require much changes:

• To stop an ongoing celery task, we can send a control signal (TERM) (ref.) but it requires the ID for the particular task. So need a way to keep track of all celery task IDs for a particular job.
• An endpoint /api/kill_analysis/{job_id} with appropriate permission enforcing i.e. If X created the job then only the admin and X should be allowed to kill it.

Checklist:

• API endpoint
• Button in frontend (IntelOwl-ng)
• member function and CLI command in pyintelowl.

[eshaan7]

So need a way to keep track of all celery task IDs for a particular job.

Could use django's low level cache API to store the task IDs (or just the celery task object if it's serializable?) for a particular job or set our own task_id.

[sp35]

Hi, I would like to work on this issue.

[eshaan7]

The API endpoint has been added, thanks to @sp35. Now we need to add support for this to pyintelowl as well as IntelOwl-ng.

[eshaan7]

We could add a kill button to the job (analysis) result page. This button when invoked should stop all ongoing celery tasks (i.e. all running analyzers) for the particular job. The status of such job should then be marked failed or killed.

Tracking in IntelOwl-ng as intelowlproject/IntelOwl-ng#70.

[eshaan7]

New member function def kill_running_job(job_id: int) -> bool in pyintelowl.

Tracking in pyintelowl as intelowlproject/pyintelowl#67.

EDIT: also corresponding CLI command.