Add Latency Filter Attribute To Delete job if timeout is exceeded

[Excommunicated]

This is for Issue #105
Added attribute based on the discussion in the listed issue..
Please let me know if there is anything I have misinterpreted.

[odinserj]

Awesome! And sorry for huge delay. I'll incorporate this changes to 1.1.2.

[odinserj]

Oh, can you merge these changes with the upstream (I can't automatically merge this PR)?

[Excommunicated]

Merged upstream and CI build is passing.

[ses4j]

Nice work, and thanks - we needed this feature. One small question, though: "Latency" doesn't seem like the right word in the DeleteOnLatencyTimeoutAttribute. What do you think about something such as ProcessingTimeLimit? Processing describes the state it is in during this period, tying it into the verbiage from the frontend website.

[ses4j]

Also, not to complicate things, but you may want to think ahead to implementing soft and hard timeouts. A soft timeout would set the cancellation token to "cancel yourself!" and a hard timeout would terminate the worker. See, for example, celery's design: http://celery.readthedocs.org/en/latest/userguide/workers.html#time-limits

[Excommunicated]

I'll let Sergey make the decision on the naming of the attribute. As for your other ideas, Once this gets merged, you should submit another issue so discussion can happen. It sounds like a good idea though..

[keith-kurak]

Are there any plans to merge in this change? This sounds really useful. Or is there another way to fail a job after a specified timespan has been exceeded?

[odinserj]

Okay, this filter was finally merged into Hangfire under the LatencyTimeoutAttribute name. It really acts on the basis of latency, and applied to a background job that is going to be processed right now (and this is not the processing time limit), so the naming is fine.

@Excommunicated, thank you for this, and for patience 👍 ❤️ I've merged it using git command line, preserving your authorship (03a3414), so I'm closing this PR instead of merging it.