A method and apparatus for enabling tracing operations in clusters of servers. An embodiment of a method includes receiving a first message at a first managed bean on a first server node of a cluster of servers, and enabling a performance trace of a computer application on the first server node in response to the first message, where the computer application is running on a plurality of server nodes of the cluster of servers, including the first server node and a second server node. The method further provides for sending a second message from the first managed bean on the first server node to a second managed bean on the second server node, and enabling the performance trace of the computer application on the second server node in response to receipt of the second message.