Should be more resilient against remote webdriver failures

[rtyler]

Sporadically (~0.5%) tests will fail with an unhandled error coming out of the remote web driver, like this traceback.

In these cases it's probably best to try to re-run the test case or mark the test as skipped

[rtyler]

Looks like this is occurring a closer to 1% for @lookout tests. I'll have to bump the priority of this.

[rtyler]

Starting work on this now.

(This message brought to you by Hub board)

[rtyler]

Since we're already using an Around block for Sauce tagged cucumber scenarios, I think it might be feasible to use the Ruby retry to re-run the scenario (block.call) up to 2-3 times.