From an infrastructure monitoring perspective, ensuring that Oracle WebLogic Server and all its underlying components are functioning should be your primary concern. In this section, we describe the monitoring of some core areas that have the largest influence on Oracle SOA Suite, namely managed servers, JVM, JMS destinations, and data sources.
As long as your managed servers are reported as healthy, there is usually not much to worry about. A warning state does not necessarily indicate that the managed server is unresponsive, but the cause of the warning should be investigated nonetheless. One of the key issues to managed server monitoring is ensuring the appropriate monitoring of threads.
To view the state of the managed servers, perform the following steps:
Ensure that the managed servers are running and are in a healthy state as designated by the green checkbox with OK. In Chapter 9, Troubleshooting the Oracle SOA Suite 12c Infrastructure, we have discussed the approaches to troubleshooting warnings or failed managed server states.
Although there are too many areas surrounding JVM monitoring to describe here, three of the more important ones include ensuring that the heap allocated on the JVM is appropriately sized (that is, comparing heap versus non-heap usage), that there is not excessive garbage collection, and that JVM thread performance is acceptable. Perform the following steps for the garbage collection information:
Oracle WebLogic Server 12c also provides a dashboard that provides real-time monitoring of many metrics, including the JVM runtime heap. This is helpful for reviewing the server while there is a heavy load on the system as it allows you to view the current and free heap size.
Go to http://adminhost:7001/console/dashboard
. Select JVM Runtime Heap and click on the start button. Heap Size Current and Heap Free Current are graphically displayed, as shown in Figure 6.22.
If the free heap hovers around zero for a considerable time, this is an indication that the heap size may be configured too small. If repeated and frequent garbage collections occur without much memory being freed up, additional JVM monitoring may be required at that point.
From the Oracle WebLogic Server Monitoring Dashboard, the Thread Pool Runtime can also be monitored in real time. The key is to monitor the Hogging Thread Count and the Pending User Request Count. In a low-usage environment, these should hover around zero. From the Oracle WebLogic Administration Console, navigate to Servers | [wls_soa1] | Monitoring | Threads to view similar information. Various thread pool metrics are displayed there. Everything from Active Execute Threads to Hogging Thread Count is shown on this page.
The Throughput shown on this page is a single value that denotes the mean number of requests completed per second. The higher this value, the better. But the thread pool changes its size automatically to maximize the throughput, so in normal cases, there is nothing you need to do aside from monitoring it to understand the behavior of your server under different types of load.
More often than not, many of the integrations that run on top of Oracle SOA Suite 12c may leverage JMS destinations to support asynchronous integrations. These destinations can be JMS queues (for point-to-point integrations) or JMS topics (for publish-subscribe integrations). Oracle WebLogic Server 12c provides a way to easily create these JMS destinations that become accessible via a JNDI lookup in the code.
As an administrator, you must be aware of all queues and topics created, as there are many reasons why you would want to monitor them:
To monitor your JMS destinations, perform the following steps:
For data source monitoring, usually all that is needed is the connection pool configuration to be valid (that is, there are no connectivity issues to the database) and the number of active connections does not approach or exceed the maximum configured connections. There are cases, such as where the time it takes to establish a connection is long (refer to the Connection Delay Time column), but these are not common and usually appear in inadequately sized environments.
To check the JDBC Data Source Runtime Statistics, go through the following steps:
mds-soa
).In cases where the database may not be accessible or is down, the database password used for the connection pool has had its password reset, network-related issues occur, or data source-related errors begin appearing in the logs, it may be worth testing to ensure that the data source is working properly by performing the following steps:
mds-soa
).Test of [mds-soa] on server [wls_soa1] was successful.
3.145.175.253