Job-level max_run_time not being taken into account after worker crash.

[tomyyn]

Open this issue after not getting a response from this comment: #832 (comment)

Currently, we are implementing job-level max_run_time for our application. Normally, if the process doesn't die, it works as expected. However, When restarting after a worker process crash, the job-level max_run_time is ignored, and only Worker.max_run_time is taken into cosideration.

As we run in a containerised environment, sometimes, containers that may have initiated processing a job are killed due to scaling down, but job remains in locked state until default max_run_time (Worker.max_run_time).

Is there any solution for this? If not, is it in the plans to implement it?