As a Site Reliability Engineer supporting a Go application running on Google Kubernetes Engine (GKE) in production, you observe that the application restarts after approximately 15 minutes of runtime following a new release. After enabling Cloud Profiler, you notice the heap usage grows continuously until the restart occurs. What steps should you take to address this issue?