Killing jobs adversely affects qstat

Killing jobs, appears to slow down qstat.
Usually qstat takes fractions of a second, but after killing of jobs qstat can take a long time to execute more than 18s. We have also seen it take more than 1minute to execute and cause
“pbs_iff: cannot connect to host”
“pbs_iff: all reserved ports in use”
Warnings, after killing jobs.
Not really sure how to debug exactly what is happening, any help would be appreciated. Thanks

This appear to be related to having large job history.
Users then refreshing the GUI impatiently wanting to see the job finish, upset qstat.

Consider it closed. Thanks