A good number of connections drop during a rolling update of Kubernetes pod running a Dropwizard app.

I updated the Dropwizard shutdownGracePeriod and Kubernetes terminationGracePeriodSeconds values to 60s and 90s respectively (defaults being 30s and 60s resp.) and I didn't see any difference in the results. Any suggestions on this as I'm trying to achieve minimal downtime during the upgrade.