Db2 for z/OS and its ecosystem

 View Only

Monitoring and controlling connections and threads with Db2ZAI 1.3.0

By Eric Radzinski posted Tue June 30, 2020 04:17 PM

By Eric Radzinski

The latest release of IBM Db2 AI for z/OS, Db2ZAI 1.3.0, which is generally available as of June 19th, comes equipped with an entirely new feature called Distributed Connection Control that's designed to ease the burden of ensuring that distributed applications are consuming connections and threads efficiently and at sustainable levels. Collectively, the Distributed Connection Control capabilities in Db2ZAI 1.3.0 can help ensure that no single application impacts the performance of your other applications by consuming an inordinate number of available connections or threads.

Db2ZAI's Distributed Connection Control feature aggregates performance statistics about your distributed workloads over time to calculate recommended thresholds for connection and thread usage. These thresholds are based on the high-water marks that Db2ZAI observes during its training process and are implemented through the use of Db2 for z/OS profiles that Db2ZAI creates for you.

When these thresholds are exceeded, Db2ZAI generates warning alerts that inform you about potential problems before they become real problems, and it generates exception alerts that actually regulate the connections and threads that are available to distributed applications when workloads exceed acceptable levels. Custom thresholds are created for each known IP address that is discovered during training to provide you with granular control at the individual IP address level. For IP addresses that are not discovered during training, which are referred to as unaccounted-for IP addresses and that can quickly generate a high number of requests for connections and threads, a common set of thresholds are created that are applied to all IP addresses of this type.

Distributed Connection Control also gives you the opportunity to prioritize your mission-critical IP addresses over less-important IP addresses so that your high-priority applications are never at risk of performance slow-downs or outages due to an insufficient number of available connections or threads.

These capabilities have been integrated into the Db2ZAI user interface in the form of the Distributed connections dashboard, which you use to establish thresholds, create and activate profiles, view and analyze the statistics about connection and thread usage within your Db2 subsystems, and monitor warning and exception alerts. For example, the Alerts tab of the Distributed connections dashboard, shown below, gives you a comprehensive view of the alerts that Db2ZAI has generated.


From this tab, you can drill down into a particular alert to display detailed information about the current state of an IP address compared to the state of that IP address when thresholds were first established for it. This information is referred to as a scorecard statistics comparison and is provided for dozens of individual connection-and thread-related statistics. For example, the following set of statistics provide information about why threads are being prevented from being pooled:


You can also display a graphical comparison of these statistics across all members of a data sharing group with the click of a button.

By organizing all of these statistics into coherent categories, Db2ZAI makes it easy for you to recognize anomalies with connection and thread usage and to perform detailed analyses of both potential and actual problems.

We’re excited to share these details about what we think is an important and valuable tool for Db2 for z/OS DBAs, and we hope you take a closer look at what Db2ZAI 1.3.0 has to offer.

Eric Radzinski is an Information Developer for Db2 for z/OS.

Sign in and subscribe to always get the latest news about Db2 for z/OS from the IBM lab: http://ibm.biz/db2znews-subscribe