Previously when doing bugzilla queries we could pass a minimal column_list parameter which was a big speed boost.

Using the python-bugzilla library, I was previously doing

        bzapi.query({
            # 'query_format' only needed with rhbugzilla4
            'query_format': 'advanced',
            'emailtype1': 'substring',
            'emailassigned_to1': True,
            'email1': me,
            'column_list': ['bug_id'],
        })

Would complete in 3-5 seconds. Now it's taking over a minute.

Here's a comment from /usr/bin/bugzilla about why it's a useful optimization, if it's any help:

        #to optimize speed and reduce network traffic through lookups, we
        #specifically tell bugzilla the exact data we want.  This allows us
        #make one call to output the data, rather than have a follow-on
        #getbug() to grab more info than what the default has.
        #Testing has shown this to be a _huge_ time saver.  Unfortunately,
        #this list and the output format have to be in _sync_.  Otherwise,
        #you lose speed by doing a look-up for each new output element.