This caused the server to be packed which resulted in failure. For example, not only the errors but also the well performing messages were monitored and saved. The review showed several performance issues. Together with Integration Matters we have reviewed the existing BWPM product (a product of Integration Matters sold by TIBCO). To avoid such situaties the energy provider was looking for a monitoring solution that would monitor the whole end-to-end chain, to quickly locate problems and even to prevent them. A result of such a problem can be that the client portal gets unreachable and unsatisfied customers start calling. Whenever problems were not found this had large consequences for the business. The chance that you miss something in a logfile is high. That took time and the origin of the problem was found too late, or even not at all. Problems were not being foundĮvery application manager then had to go through each of his logfiles to find the mistake. Whenever something went wrong people had a hard time figuring out where the problem was situated and often did not find the problem at all. The applications surrounding the BW were not being monitored. This product only showed what went in and out of the Business Works. For TIBCO BW they had Business Process Monitoring (BWPM) installed. Untill now monitoring was only performed on a application/platform level. tra file for one of the Tibco BW nodes.A Dutch energy provider was looking for a monitoring solution which would provide insight in the end-to-end processes. Per the info in PREREQUISITES AND TROUBLESHOOTING, only the port number in monitor.xml must be changed - to the value used for the JMX port in the. Copy monitor.xml (with the port number changed) and the jar file for the monitor into that directory, and restart the machine agent. The default metric tree prefix isīelow that will be a process name, and under that count and duration metrics.Ĭreate a directory under monitors for the extension. Per the info in USE CASE: count and duration for Tibco proesses, by process. tra file for every process to be monitored. tra file option Jmx.Enabled=true must be set in the. Tibco processes must be run with the standard JMX localhost command-line options: Info-level logging is not great for troubleshooting, but debug-level logging should give reasonable information on how the JMX connection is working, and on what stats objects are being returned from JMX. The other values should work fine with the defaults, only change if there is a use case reason to do so. The port is the only value in monitor.xml that MUST be changed. This extension levarages localhost JMX, so the machine agent running this extension should be on the same OS instance as a Tibco process with JMX enabled. That port number must be configured in monitor.xml for this extension to work. When enabling JMX in Tibco BW, you select the JMX port in the '.tra' file of your Tibco deployment.

% mvn -Dtibhawk=/Users/paulbunyan/tibco/bw6/hawk/5.1 -Dappd-ma=/Users/paulbunyan/machine-agent clean package For example, if you have Tibco installed in a directory called /Users/paulbunyan/tibco and an AppDynamics machine agent installed at /Users/paulbunyan/machine-agent, then you build using: Variables are used in the build files to resolve these dependencies. Also, the AppDynamics machine agent is a dependency. To build you must have Tibco dependencies available, they are not available through Maven. Process counts are important because in a BW environment there is usually a fixed thread pool available for running process jobs - high counts mean that jobs are sitting around waiting for an available thread. This extension monitors Tibco BW processes and their sub-processes ('Jobs'), giving count and duration metrics for each. Do not try to build the hawk console monitor in the github repo, it will not build and is as of now just a template. Tibco BW must be run with JMX enabled for this extension to work.Īs of this writing the Hawk Console monitoring capability is not functional, so only the JMX monitor is usable.

Tibco bw process monitor