Average Response Time for Operations
The Average Response Times for Operations viewer Layout for Charts provides a horizontal Bar element visualizer component to display the average response time that is calculated across all instances of specific methods or commands of the same type that are detected in a set of trace results, that is, for protocols that make use of request/response pairs. For example, an HTTP GET or POST method is a request message that is sent by a client to a web server, where the client awaits a response from the web server. When a response is received, Message Analyzer encapsulates the request and response messages in an Operation node that displays as a top-level message row in the Analysis Grid viewer. The Average Response Time is calculated as the average time it took for all request messages in a set of trace results to receive the first server response.
Understanding the Average Response Time
The ResponseTime annotation from Field Chooser enables you to measure the difference between the time that a Request message was sent by the client to a particular server and the first Response message received by the client from the server. Therefore, the Average ResponseTime indicated by a bar element for a specific method or command provides an average value that tells you how long a particular server is taking to reply to requests. Because the bar graph data is sorted from the longest to the shortest average response times, you can quickly ascertain from the upper rows of data which requests are taking the longest time to get a response for a particular method or command, possibly pointing to server performance problems. The Average Response Time that is associated with each method or command is also specified on the right side of the bar graph as a percentage value that is relative to the total response time for all Operation messages in a trace. This percentage value can also provide a quick indication of which server/s are taking the longest time to respond. With this information, you can focus your troubleshooting efforts on performance of the server that is sending response messages to the client.
To determine which server may be having performance problems, you might be able to isolate that information through filtering. For example, for HTTP GET methods that are reporting a long average response time, you might apply a view Filter to the Analysis Grid viewer such as HTTP.Method == "GET" to isolate the Operations that contain the HTTP requests. You can then add the ResponseTime annotation from the GlobalAnnotations node in Field Chooser as a new column in the Analysis Grid viewer, from where you can correlate the high response time values with the IP address of the web server in the Destination column. You might perform a similar filtering operation for an SMB2.ComNegotiate request message to a file server.