best practises for incident rule creation
Does Compuware provide any best practises or suggestions regarding what middleware systems resources (i.e., JVM resources, WebLogic resources) should be monitored via custom incident rules? In the past, when we had only Server Vantage to monitor the health of our applications, we defined a large number of alert rules on JVM and WebLogic resources. We could, of course, implement similar incident rules using the out-of-box JVM and WebLogic metrics that dynaTrace makes available, but I suspect that this might be overkill. My thinking in stating this is that, using dynaTrace, we have the ability to implement incident rules on business transaction performance and errors as well as API performance and errors To some extent I believe that this "top down" monitoring likely obviates the need for some of the JVM and WebLogic resource alerting that we configured in Server Vantage. Does anyone have suggestions regarding what a reasonable set of JVM and WebLogic alert rules might be, or does Compuware provide any best practises in this regard?
Hi Jeff
I think you are spot on. Most of the alerting with dynaTrace should be around performance degradations on the application and end user. For that reason we also have our dynamica baselining on business transactions that you can use
For System Health Monitoring we also provide out-of-the-box monitoring of process and host health metrics. I am sure you have seen the infrastructure view and the process and host health dashboards. We automatically look at several key performance dimensions, e.g: cpu, memory, disk,network (on the host) and GC (on the process). There are out-of-the-box incidents that trigger in case your system becomes unhealthy on one of these dimensions.
On top of that you can of course create your own Incidents on dimensions that are important for you, e.g: throughput, connection pool utilzation, ...
hope this helps
