The correct option is Reconfigure the monitoring agent to collect only from the active service.

Reconfiguring the agent is the precise and least disruptive action because it stops the agent from collecting irrelevant data while keeping monitoring for the active service intact. Adjusting the agent's target filters or service discovery settings lets you restore accurate metrics and alerts without removing monitoring from other hosts.

Export and archive the collected logs to Cloud Storage before investigation is not the best immediate action because exporting does not stop the ongoing collection or correct the misconfiguration. It only preserves data and can add unnecessary storage cost while the root cause remains.

Leave the agent running without making changes is incorrect because it allows continued collection of irrelevant data which leads to noisy metrics, potential false alerts, and extra cost. Passive acceptance does not resolve the problem.

Stop and uninstall the monitoring tool from all hosts is overly broad and harmful because it removes visibility for all services running on those hosts. Uninstalling is disproportionate when a configuration change will fix the issue with far less risk.

Full AWS Practitioner Certification Question