For some time now, Dynatrace has provided you the ability to automatically detect when the unavailability of a specific process results in a problem (for example, increased failure rate or outages of your website). Recently we announced that you can receive notifications when a process within a specific process group becomes unavailable—even if no related customer impact is detected.
We’ve now added another enhancement that’s particularly useful for cluster monitoring: setting a minimum number of process instances for a process group.
A “process group” is a group of processes that each do the same thing (for example, a group of identical Apache Tomcat web server processes that run concurrently on separate hosts). Each process in a process group (aka, a “cluster”) serves the same type of request. Typically, a load balancer is run in front of such a cluster.
In such scenarios, it makes sense to define a minimum number of process instances for each process group—the minimum number of processes that comprise a healthy cluster.
To define a minimum number of instances for a process group
- Select Technologies from the navigation menu.
- Select the tile of the process-group technology you want to configure (for example, Apache Tomcat).
The related process groups are then displayed at the bottom of the page.
- Select the process group you want to configure.
The process group view expands.
- Click the Process group details button to open the process group page.
Instances of the selected process group are then displayed at the bottom of the page (for example, the four instances of the
CustomerFrontendprocess group shown below).
- Click the Edit button (…).
- In addition to the pre-existing Impact-only and Immediate anomaly-detection options, a new Threshold option is now available.
- Select Threshold from the drop list.
- The minimum number of processes for the process group defaults to the current number of process instances in the process group (4 in the example below). Set the value to the appropriate minimum number of process instances required for a healthy system.
Once defined, you will be alerted going forward if the number of processes in the process group falls below the minimum that you’ve set. Once the defined minimum number of instances is met, the problem is closed automatically.