Add redis pooling and worker refactor

[nerdalert]

• Added a redis pool to resolve #165
• Some refactoring to create an instance of a new worker struct for each new job.
• Added job ID to results for better debugability and tracking.
• Added a precheck error reporting. One thing of note is if no yaml files were submitted, chat would run without errors resulting in an empty report. This checks if no yaml files are in the diff and exits the runPrecheck() method early to avoid unnecessary processing since precheck could be a default action on PRs. It also avoids an empty report.

[nerdalert]

@russellb Validated a few concurrent jobs over two PRs (the error intended on job 10 in the screencap). The worker redis client no longer bombs when running more than one job at once.

[russellb]

The worker redis client no longer bombs when running more than one job at once.

Each worker should only be running one job at a time, not doing them in parallel. Or do you just mean more than one job in the queue at the same time?

[nerdalert]

@russellb if you queued up two jobs concurrently, the redis connection bombed because the connection is closed by the previous job so it had a conn that was closed it was using to try and query redis. My guess is somewhere in the go routines it was getting triggered even though it wasn't running the job. Jobs are still sequential. You can see in the following a two minute gap between results. A screenshot was cleaner then logs since we can't use pub gists 😬

[nerdalert]

Also validated across two workers with two jobs. They each ran one of the jobs. 🎉