Ship worker logs to datadog

[hackartisan]

The worker machine doesn't appear to ship logs to datadog, and it also appears to truncate when the clock rolls over to a new day. We should have more log retention either on the box or in datadog.

sudden priority justification

We don't currently have visibility into the timing of our jobs. Lyrasis is observing api hits that concern them, and asking us about it.

Acceptance Criteria

• There are logs for the production worker box in Datadog.

First Step

Go see how we're getting logs for the web boxes for pulfalight in princeton_ansible, and get it working that way for the worker box.