Basic retry on standard failure

[tpitale]

• [x] callback on job module
• [x] trigger retry_at on standard failure
• [x] proceed with normal failure procedure on :noop return
• [x] include rihanna_internal_meta when querying/inserting jobs
• [x] store next due_at datetime and increment attempts in meta jsonb
• [x] test dispatcher
• [x] test Job
• [x] docs on usage, including about Jason requirement for Postgrex.

Future features could include:

• retry_in callback, which just returns seconds until next retry
• automatic retry handling with a retries limit option
• alternative backoffs: exponential, linear, immediate, etc

Related to #32

[tpitale]

@lpil @samphilipd Can I get a review of this?

[tpitale]

This will be cleaner after merging #65.

[samsondav]

Will wait until #65 is in :+1:

[tpitale]

@samphilipd I rebased this off master, so it should be safe to merge after you review.

As this doesn't require a breaking change and a V2 release, I figured we might want to add it before the priority changes.

[tpitale]

@samphilipd Let me know what you think of the docs. Thanks!

[tpitale]

🎉🎉🎉🎉

Thanks @samphilipd