Clusters out of sync

I have a multi cluster setup (primary and standby) after a while (say couple of days to months usage) and few failovers from primary to secondary and vice versa
i see that the two clusters get out of sync.
When i do
tctl -ns mynamespace la --op | wc

i get about 300 workflows in actvie

while i see that my secondary has about 600 or 700 workflows

in the due course i have upgraded my cluster from 1.14.x to 1.15.0 to 1.15.2
not sure when and how the clusters get out of sync.

Do you have workers in both primary and standby?

i have temportal fornt end, workers ,hostory and matcher both in primary and standby

However, there are no workflow/ activity processor/client-sdk
in standby