Change Details

Currently, the hide numbers are only available through a hive query (see https://phabricator.wikimedia.org/T90635#1063926), which takes *hours*. Part of the issue is that we're parsing the URL many times for each record. Instead, we should be parsing and storing this data into a (public) schema with at most a 15 minute lag behind real-time, with instantaneous query access and appropriate indexes.