Elevated innodb_row_lock_time in version 10.11.7

Hello,

I am upgrading from mariadb version 10.4.33 to 10.11.7. To do a performance test, I took two read replicas from my fleet and upgraded one to the new version, then started serving traffic to both.

I monitored some metrics during this experiment and saw the following:
No increase in latency
no increase in replication lag
2-3x increase in innodb_row_lock_time and innodb_row_lock_waits

I am aware these are cumulative metrics. however, to measure them, i am checking the value every few seconds and then emitting the delta. However, my innodb_lock_wait_time_avg has gone down on the new version. Additionally, handler_rollback metric is elevated on new version as well.

Is there a known reason for the increase in this metrics? are theycalculated differently? Any other things I should look into?