You have an application that runs in Google Kubernetes Engine (GKE). Over the last 2 weeks, customers have reported that a specific part of the application returns errors very frequently. Currently, there is no logging or monitoring solution enabled on your GKE cluster. You need to diagnose the problem to understand why customers are encountering errors. You have not been able to replicate the issue yourself and want to use minimal disruption to the application to perform this diagnosis. What should you do to enable efficient troubleshooting?