As a Site Reliability Engineer, you oversee an application written in Go operating on Google Kubernetes Engine (GKE) in production. Following the release of a new version, you observe that the application runs for approximately 15 minutes before restarting. Upon integrating Cloud Profiler, you observe that heap usage consistently increases until the application restarts. How should you proceed?