Worked on the redpanda-operator repository to enhance reliability during the post-installation process by developing a configurable feature in the Helm chart. This feature allows users to exclude the post-install job pod from the Redpanda Service selector, addressing the issue of stale endpoints that could disrupt client connections. The implementation involved modifying Kubernetes labels so the post-install pod no longer matches the service selector, with the option defaulting to off to minimize operational risk. The work was carried out using Go, Helm, and Kubernetes, focusing on improving connection reliability and reducing support overhead without introducing breaking changes to existing deployments.