A Java application server is running on a host monitored by the Instana agent. Host-level metrics such as CPU, memory, and disk I/O appear correctly in the Instana UI. However, JVM-specific metrics including heap usage, garbage collection activity, and thread counts are completely absent. The agent.log shows no Java sensor errors. What should the administrator investigate next?

Host-level metrics come from the host sensor and their presence confirms the agent is running and connected to the backend. Missing JVM metrics in isolation indicate the Java sensor is not successfully attaching to or instrumenting the JVM process. Common causes include a missing -javaagent JVM startup flag, incorrect auto-attach configuration, or insufficient OS-level permissions for the Instana agent user to attach to the target JVM process. The correct answer is 'Verify that the Instana Java sensor is correctly attached to the target JVM via the -javaagent startup parameter or auto-attach configuration, and that the agent process has sufficient OS-level permission to instrument the JVM'. The incorrect options — 'Increase the Instana agent JVM heap allocation to allow the Java sensor sufficient memory to execute', 'Restart the Instana backend to trigger a fresh auto-discovery scan for the Java process', 'Add the application server to the Instana infrastructure entity filter to enable JVM-level metric collection' — are wrong because they do not align with IBM Instana's architecture or recommended practices for this scenario. Understanding this concept is essential for the Domain 6: Troubleshooting domain of the IBM Instana Observability certification.