fix(rq): Only capture exception if RQ job has failed (ignore retries)

[BobReid]

Fixes #1074

Sentry reports all errors from RQ even if it is still retrying the job.

This changes checks to ensure that retries have been exhausted, ie. the job has failed and will be moved to the failed jobs queue, before capturing the exception.

Changes:

integrations/rq.py

• only call _capture exception if job.is_failed is true, indicating it is moving to the failed jobs registry
• add ignore_logger('rq.worker') so the exception does not get captured via the logging integration

tests/integrations/rq/test_rq.py

• Add a failing test case for redis version 1.5 and above (where Retry was introduced) to ensure only one exception is captured in the event of a retry

[BobReid]

Addresses #1074

[untitaker]

I cannot merge master into your branch but I've manually verified that this works.