Old active connections?

Hello group,

We run postgresql 10.3 for a python django app with gunicorn on nginx with
django version 1.9.5.

Recently, we started noticing there are many active connections from the
django app server that are more than 1 week old still showing in
pg_stat_activity.

Even though the django server has been stopped (all processes killed), the
active connections still persist. All of these connections are UPDATE
queries that look pretty normal.

Does anyone know the reasons they could be there? What could have caused
them being still active?

This has become an issue as we started getting "FATAL: remaining
connection slots are reserved for non-replication superuser connections"

Thanks,
-hans