add error rescue for cron start method

[carlad]

There's an issue where any error in the start method is not rescued, and will bubble up to the start_all method, causing all subsequent crons to be aborted.

This change has been deployed to staging and everything seems to be working.

[joecorcoran]

Looks good. Only thing I would add is that it would be more useful to log the error with whatever logger we usually use in this app, rather than using puts.

[rkh]

Looks good. Only thing I would add is that it would be more useful to log the error with whatever logger we usually use in this app, rather than using puts.

I think that's accessible via Travis.logger. Also, we might want to track exceptions in Sentry as well, so they don't get lost.