What Is Communilytics? : A community analytics funnel in practice
via slideshare.net
Explore Posterous
Search posterous
Search all posts and users. Type a name, type a favorite song title, whatever! See what comes up.
More posterous blogs
Here are posterous posts filed under monitoring...
Monitoring Tomcat with Java VisualVM
One of the best kept secrets that's bundled with your Java 6 JDK is VisualVM. VisualVM is an absolutely fantastic, free monitoring tool for Java that you may not realize is right under your nose.
In a nutshell, when you fire up VisualVM it provides you with a ton of monitoring tools for everything running on the JVM. By default VisualVM will monitor the VM from which it's launched, so if for example you launch VisualVM from the bin directory of jdk1.6.0_17, then anything using that JVM will show up as a process in VisualVM that can be monitored. Note that in some cases certain processes will not appear in VisualVM, which is the real point of this post; I'll get to that in a moment. I'm attaching a few screenshots to this post. The first shows what VisualVM looks like when I fired it up just now on my Linux laptop, the second shows a snapshot of the main monitor screen (in this case I'm monitoring Tomcat), and the third shows a snapshot of the thread monitoring screen (note the Open BlueDragon threads!). VisualVM can also take and load snapshots so you really hone in on problems at the VM level quite easily. One major point I'd like to make is that VisualVM will monitor any Java application running on the JVM. So in the CFML world this means if you have OpenBD, Railo, or ColdFusion running, they can all be monitored quite nicely using VisualVM. VisualVM can also monitor remote JVMs via JMX (Java Management Extensions), so if you're having trouble on a remote server and want to see what's going on, as long as the JMX ports are open and accessible you can launch VisualVM from your local machine and connect to the remote VM. Note that monitoring is a very lightweight process so VisualVM can be used to monitor production servers with virtual no impact. Visual VM also does profiling, however, which is a much more heavyweight process. It provides a huge amount of useful information, but should be used only when absolutely needed on production servers since it will have a noticeable impact on performance. Now to the real point of this post. When I launched VisualVM on a Windows 2003 server today I was surprised that Tomcat didn't show up as a process running under the VM. Turns out that if you install Tomcat as a service, even if it's running under the same user account that you used to launch VisualVM, Tomcat won't appear by default in the VisualVM process list. Luckily it's easy enough to resolve. Simply open the Tomcat Configuration application and add the following in the Java Options box on the Java tab:-Dcom.sun.management.jmxremote.port=8086 -Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=falseAfter making this change you do have to restart Tomcat.This will enable JMX in Tomcat and allow VisualVM to connect to it. You can choose any port you like, and note that if you want to use SSL or authentication you would set those options to true. I haven't personally messed with authentication so I'm not sure what that authenticates against, but know that if you want to have JMX available on a production system that you can secure it this way, or of course through firewall rules. With JMX enabled in Tomcat you then go into VisualVM, add a new JMX connection, and point it to localhost:8086 (or whatever port you set JMX to run on). That's it--you're now monitoring Tomcat!VisualVM is a great, free tool that you likely already have on your machine, so you really owe it to yourself to check it out.
Did the FAO's World Summit go far enough? "To my regret..."
(excerpt from CropBiotech Update /; November 20, 2009)
The head of the United Nations Food and Agriculture Organization Jacques Diouf concluded the World Food Summit by affirming that it is possible to free the world from hunger but called for a "move from words to actions." The World Food Summit, held earlier this week in Rome, brought together 60 heads of state and 191 ministers from 182 countries and the European community. Diouf said that the Summit produced four important commitments. These include:
* a firm pledge to renew efforts to achieve the First Millennium Development Goals of halving hunger by 2015,* a pledge to improve international coordination and the governance of food security through a profound reform of FAO's Committee on World Food Security (CFS),
* a promise to reverse the downward trend in funding for agriculture, food security and rural development in developing countries, and
* a decision to promote new investments in agricultural production and productivity in the third world. However several critics questioned the Summit's outcome. All but one of the G8 leaders, Italy's Silvio Berlusconi, failed to attend the Summit. Even Diouf himself acknowledged that the Summit "didn't go as far" as he had hoped. "To my regret the official Declaration adopted by the Summit this past Monday contains neither measurable targets nor specific deadlines which would have made it easier to monitor implementation," Diouf said. The UN Agency said that world hunger can be eradicated if rich countries hand over USD 44 billion a year. Read http://www.fao.org/news/story/en/item/37465/icode/ for more information. The transcript of Diouf's closing speech is available athttp://www.fao.org/fileadmin/user_upload/newsroom/docs/dgfinalspeech.pdf
javamelody: monitoring of JavaEE applications
http://code.google.com/p/javamelody/
* make decisions when trends are bad, before problems become too
serious
* optimize based on the more limiting response times
* find the root causes of response times
* verify the real improvement after optimizations It includes summary charts showing the evolution over time of the following indicators: * Number of executions, mean execution times and percentage of
errors of http requests, sql requests or methods of business
façades (if EJB3 or if Spring)
* Java memory
* Java CPU
* Number of user sessions
* Number of jdbc connections These charts can be viewed on the current day, week, month or year. JavaMelody includes statistics of predefined counters (as of today http requests, sql requests and methods of business façades if EJB3 or if Spring) with, for each counter : * A summary indicating the overall number of executions, the average
execution time, the cpu time and the percentage of errors.
* And the percentage of time spent in the requests for which the
average time exceeds a configurable threshold.
* And the complete list of requests, aggregated without dynamic
parameters with, for each, the number of executions, the mean
execution time, the mean cpu time, the percentage of errors and an
evolution chart of execution time over time.
* Furthermore, each http request indicates the size of the flow
response, the mean number of sql executions and the mean sql time. It also includes statistics on http errors, on warnings and errors in logs and on data caches if ehcache. An optional and independent collect server may be used if necessary to unload the application of storage management, and of reports generation and to centralize the data of clustered applications or of several applications.
Is a single social media monitoring platform enough?
A great argument for using multiple monitoring tools.
