Your organization runs a web application on Google Kubernetes Engine (GKE). During periods of high traffic, the application's CPU usage frequently spikes above 75% but typically stabilizes within 2 minutes. The current alerting policy sends notifications whenever CPU usage crosses 75%, resulting in too many false positives. You need to configure an alerting policy that only triggers if CPU usage exceeds 75% for at least 3 consecutive minutes. What should you do?