URL auto learning

URL auto-learning enables you to define the set of URLs appearing in per-URL reporting statistics without the need to manually define each URL. URLs that are found frequently are learned and reported.

The NAM Probe uses a URL auto-learning algorithm to choose and report the most popular URLs observer in monitored traffic.

The NAM Probe uses a pool of auto discovered URLs. It consists of members and candidates. The auto-learning algorithm aggregates the loads of all URLs for all monitored servers. The server IP address, port or any other attribute of the server is not used by the auto-learning algorithm. If a URL becomes the member of the pool, it is reported for all servers, regardless the activity on individual servers.

URLs are inserted into the list of candidates and the list of members and moved between the lists according to the configuration parameters set in the Advanced setting section of the URL Auto-learning screens in the NAM Console.

The NAM Probe removes the URLs from the member list every specified interval, as controlled by the Cleanup interval property.

Use the Percentage of new URLs property to control the portion of the members pool to be freed at the beginning of each interval and reserved for new highly active URLs. A candidate URL becomes a pool member if it is more popular than the portion of members defined by the Page loads threshold property.

If you want to report URLs from all servers in the farm, regardless of the individual host name, you can deselect Use host name to exclude host names from the URL auto-learning algorithm.

Enable the Slow page weight property to ensure that slow operation URLs with high volume loads are included in the auto-learning algorithm.

To avoid a situation when no URLs are reported for software services with very little traffic, you can set the rules at the any software service level. In this way, you create a software service level pool with lower limits, making sure its URLs are reported.You can create separate pools within a single software service based on a number of servers. This way, you ensure the URLs monitored on a server with a lower traffic do not have to compete with URLs from a much larger server in terms of volume. You achieve this by assigning servers to groups within a single software service which translates to separate pools. Use the Group name option in the Service Details dialog. For more information, see User-defined software services.

You can also create separate monitoring pools for default software services based on HTTP and HTTP express analyzers, both globally, and the software service level, by enabling URL auto-learning for default software services.