Thanks to https://jira.nypl.org/browse/SIMPLY-529 we now track how long scripts, monitors, and coverage providers take to run. If there's an exception that crashes the service, we track the exception. However, we don't keep track of what actually happened while the service was running.

We have a field to track this, timestamps.achievements, but we don't use it for anything. Now's the time!

Here's how I have it planned out:

• Create a TimestampInfo class in metadata_layer, which abstracts away the details of a Timestamp object.
• Change the parent classes Script, CoverageProvider, and Monitor to expect a TimestampInfo when they call the "do the thing" method overridden by the subclass. Upon receipt of a TimestampInfo, save it to the database.
• Change Monitor and CoverageProvider to do a sensible default thing with their TimestampInfo – e.g. CoverageProvider should tally the number of successes, transient failures, and permanent failures it encountered.
• Change individual scripts to report on their progress depending on what the script does.