Scenario: An application in Google Kubernetes Engine (GKE) has been experiencing frequent errors in a specific part over the last 2 weeks, as reported by customers. There is no logging or monitoring solution enabled on the GKE cluster. Question: What steps should be taken to diagnose the issue without causing significant disruption to the application, considering the inability to replicate the problem?