ring server`  #2195

add a try/catch block to the code which introspects a Mongo collection so that if an unexpected exception is thrown we can catch it and have a nice stack trace while not affecting the sync of other collections.

update slack api settings endpoint to handle `metabot-enabled` setting and add a new metabot-lifecycle event handler which can listen for the appropriate events (:settings-update) and trigger startup/shutdown of metabot as needed.

add a new setting for `metabot-enabled` which defaults to true and only launch metabot when it's enabled.  also, remove printlns and tidy up logging a bit.

remove `field_type` as a user presented/controllable thing.  only meaningful implication is that the list of fields for breakout becomes all fields by default now.

publish a `:database-update` event and create an event listener which calls our driver notify-database-updated method whenever we get a :database-update event.

add a new optional driver method for `notify-database-updated` which allows drivers to be informed when their databases are updated. then implement this method on the generic-sql driver and use it to actively shutdown any open connection pool on the updated database so that we can recreate it as needed.

update the connection-spec acquisition code for our generic SQL driver to unify it based on acquiring a connection for a given DATABASE and to replace the memoize/ttl with our own atom for caching db connection pools to each database.  This avoids the unnecessary recreation of connection pools every so many hours.

tidy up a few more places where Field :active true was being used and replace it with :visibility_type "retired"