If you choose Interview Web as the Performance object, you have access to the following performance counters. If you have more than one instance of Interview Web registered, you might want to monitor each instance separately.

The mrIWeb and mrIEngWS cluster monitoring average response time counters have been updated to average over the last interval rather than over the time since started. They are now PERF_AVERAGE_TIMER type counters instead of PERF_COUNTER_RAWCOUNT type counters.

Number of active threads.

Similar to the Interviewer Session Engine performance object's Active Threads counter. This is the total number of threads that handle web requests. Web requests are queued upon arrival. If a thread is available in the thread pool in the Interviewer Server Web Service, a request is removed from the queue and the available thread starts handling the request. At this time, the counter value increases by 1. After the request is handled, the counter value decreases by 1.

Average time spent handling a request.

Similar to the Interviewer Session Engine performance object's Average Response Time counter. The Interviewer Server Web Service records the total time spent handling all requests and the total number of handled requests. The counter value is the result of the total handling time divided by the total number of handled requests.

Current number of requests waiting to be handled.

When a request arrives, the counter value increases by 1. After a thread starts handling a request, the counter value decreases by 1. A sustained reading of greater than 100 indicates poor respondent response times and is the best indicator that the web server is overloaded. Adding another Interview Web Server and setting up load balancing could help increase performance. See How the cluster was tuned for more information.

Cumulative number of interviewer session engines that have been moved to the failed engines list.

For more information, see Session engine registration.

Cumulative number of interviewer session engines registered with the web tier.

For more information, see Session engine registration.

Cumulative number of interviewer session engines that have been unregistered.

For more information, see Session engine registration.

The number of cache subscriptions that failed to transfer to the database (per second).

Maximum number of requests waiting to be handled.

This is the maximum value of the Current Queued Requests counter.

Maximum time spent handling a request.

This counter uses the same data as the Average Response Time counter.

Number of failed requests.

Similar to the Interviewer Session Engine performance object's Server Failed Requests counter. When an error occurs during the web server’s handling of a request, this counter value increases by 1. In this case, the HTTP status code is typically greater than or equal to 400.

Number of requests per second.

This counter type is the PERF_COUNTER_COUNTER and shows the average number of operations completed during each second of the sample interval. For more information, see “PERF_COUNTER_COUNTER”:

The total number of handled requests, including successful and failed requests.

The total number of cache files that reside in the primary and secondary cache folders. The counter in only updated after the scan of the subscriptions completes.

The counter value is updated after the Transfer Service scans the primary and secondary cache folders, and represents the total number of cache files (*.chj) in the folders.

The primary cache folder is typically the only folder, and has a DPM site property value ofInterviewCacheUNC. The Transfer Service hosted in every session engine scans the same cache folder, resulting in the Total Cache Files counters for all engines displaying the same value. An Interview Session Engine’s counter can be monitored to track the total number of cache files residing in the cache folder.

The total number of cache subscriptions that failed to transfer to the database.