Load testing was conducted prior to migration, but unexpected issues arose once the system was in production. The autoscaler failed to properly scale the cluster, leading to deployment delays. After rolling back to a previous version of Kubernetes, stability was restored, but further attempts to upgrade introduced new challenges that required careful management of traffic between old and new systems.
