Reduce calls to detailed results endpoint

Change the way we handle catch-up messages to only fetch the detailed results URL after ALL lines of the logs have been handled, so we only get the last detailed result URL, rather than one call for everytime the graph was updated.