Adding timeouts to group and task

[mikenomitch]

During a hack week I took a stab at an initial implementation of timeouts for tasks. Pushing up this draft PR as a starting point for discussion. (Also note to anybody external to HashiCorp reading this, I'm a PM, not a full engineer, so don't get too excited just yet 😄)

What this does:

• This creates a new taskrunner hook and new allocrunner hook for timeouts.
• In the jobspec, if a new "timeout" block is set, the runners start a goroutine that waits for the deadline to pass. The allocation or task (depending on the block) is then killed at the deadline.

Open Product Questions:

• I am currently failing the allocations if the group timeout happens, and failing them at the task level if FailOnTimeout is set. It kind of feels like there should be a top level "timed out" status instead? If a task fails due to timeout, should a postStop task get run? If a preStart tails times out, should a main task get run? Do people need customization of this behavior?
• Right now this depends on a Nomad client running. Is that an issue for people?
• I think timezones need to be handled. I think you could just pass in a TZ at the jobspec level and us that? If not, default to the client's TZ? To UTC? Could we just not do this initially and always go UTC?
• This includes the functionality on service and system jobs... would that ever make sense?

Open Technical Questions:

• This is saving information about the timeout in Nomad Variables. Its sort of cool that this is inspectable by the user, but I think this should probably be "its own thing" in the Nomad state store. Any disagreement there? What's the "right way" to store this?
• Is there a better way to wait in the goroutine? I feel like this probably checks in too often and is wasting cycles?

I'm sure there are more open questions/issues (and unfortunately I've forgotten a lot since I did this over a month ago), but wanted to get the conversation going!

Fixes #1782