Discussions

News: AutoPilot optimizes WebLogic environment, detects problems early

  1. Arcturus AutoPilot optimizes a WebLogic environment, detects problems early by proactively monitoring the server for conditions that can lead to an outage. It determines when to do the root cause analysis and creation of blackbox data in the case that WebLogic gets into a fault state.

    The latest version includes significant updates to the embedded knowledge base, enhanced Blackbox, improved Advisor and Detector, real time monitoring, console mode installation, completely automated Root Cause analysis along with many other features.

    A free 15-day download is available after a simple registration.
  2. Support for WebLogic 9.1?[ Go to top ]

    Does the current version support WebLogic 9.1? From the discussions online, it only talks about WebLogic 8.0 (SP2) but there is no mention of WebLogic 9.1.

    Also, when do you plan to have support for SQL Server (instead of Oracle and MySQL) for the persistant data?

    Thanks

    Venkatt
  3. WebLogic 9.1 Support[ Go to top ]

    We are in the process of supporting 9.1 and will have a version out for WebLogic 9.1 shortly.
  4. Similar Product for WebSphere[ Go to top ]

    Do we have something similar for WebSphere
  5. We are in the process of adding support for WebSphere, Oracle & JBoss but I can't give you an ETA at this time.
  6. I noticed that AutoPilot BlackboxTM has a patent pending (as well as many of the other market features) and it seems from the outset to being similar to JXInsight's JVMInsight and Object State Graph Recording (Mememto). Is a black box basically another name for a JXInsight repository snapshot and object state graph memento?

    In the next EA build of JXInsight 4.2 we offer the ability to configure the inclusion of JVMInsight repository snapshot models into the trace memento which can be generated when an error is thrown or a performance slowdown detected.

    Regards,

    William Louth
    CTO, JInspired

    "J*EE tuning, tracing, testing, and monitoring with Insight"
    http://www.jinspired.com

    http://www.jinspired.com/products/jxinsight/contextintime.html
     Summary
    The ability to offline inspect object state and other contextual information associated with business transactions is extremely useful when performing post problem analysis on an reported application error or performance incident. This article introduces the new object field state graph recording capabilities available in the latest JXInsight 4.2 early access builds that enables application support staff to accurately perform root cause analysis.


    http://www.jinspired.com/products/jxinsight/jvminsight.html
    JVMInsight is an extensible and configurable repository inspection solution that enables the JXInsight management and monitoring console to display additional state information acquired from a server's (local and remote) repositories. Having the ability to access such detailed runtime configuration and state information from a production JVM is crucial in resolving service performance and outage problems. For developers this information provides a level of detail normally provided by a Java debugger but without the overhead. For testers the information can be stored for ongoing change & configuration management activities. With the ability to configure the automatic export of information from multiple repositories in the event of a runtime event (request failure, performance slowndown, JVM shutdown) operations and support staff have a powerful tool for post analysis of related incidents and service calls. Consolidating the performance profile, metric and tracing information with the runtime state and configuration enables remote support staff to work on resolving issues without the need to have actual physical access to the production systems.
  7. Patents, Patents, .........[ Go to top ]

    nice plug...
  8. Patents, Patents, .........[ Go to top ]

    Personally I find picking on news announcements by competitors a questionable practice, not to mention plugging in your own marketing messages into the discussions following such announcements. It does no good to the image of your company.

    I guess now guys at Arcturustech have all rights to retaliate the same way when JInspired posts a product announcement.

    Regards,

    Slava Imeshev
  9. BlackBox is nothing new[ Go to top ]

    I just finished having a look at the single documentation page related to BlackBox and it would appear to be just similar to one of the many runtime state inspection providers we supply with JXInsight via our JVMInsight technology which was previewed quite some time ago in our early access builds for JXInsight in Dec 2005. In this case this is a basic provider implementation for WLS based on JMX.

    JXInsight 4.2 ships with a JMX repository provider inspection solution for all standard JMX MBeanServers and not just WLS as well as 10 other different runtime state inspection providers.

    So what is exactly the patent about? How does this relate to the other mature APM solutions on the market that have advertised blackbox recorders for many years. Can you give me the filling reference for this patent?


    Regards,

    William Louth
    JXInsight Product Architect
    CTO, JInspired

    "J2EE tuning, testing, tracing and monitoring with Insight"
    http://www.jinspired.com
  10. You can read more about it in an article I published in WebLogic Developers Journal.

    http://wldj.sys-con.com/read/169377.htm
  11. "...auto-detects WebLogic Server failure conditions with 99.9 % accuracy"

    Is this just marketing speak? If not then please share with us how this figure has been derived as I cannot understand how one even goes about asserting this.

    With regard to the other aspects being promoted in my experience troubleshooting J2EE/CORBA performance problems over the last 10 years it has been much more effective in approaching issues from an application and external resource perspective rather than tuning of various WebLogic configuration parameters especially.

    I am not saying that you cannot make performance gains by changing a particular config value but that any proposed change must take a more holistic and application level analysis where the cost and benefit is assesed by application architects and system administrators. I am sure that basic metric sampling tools can be created that catch the low hanging fruit but for others the tool must have a greater understanding of the execution and workload patterns, JVM runtime (BEA JRockit, Sun), the IT infrastructure, security, data integrity, manageability, the business objectives (performance vs reliability vs scalability....).....

    Software performance engineering is a complex task - there are no silver bullets that magically make applications go faster by toggling a fast flag. We need to focus on creating tools that provide deeper insight into the runtime aspects of the application and platform and commicating this in the most effective and efficient manner via powerful visualizations and automated (and offline) profile inspections offering advice and important observations.

    I would like to repeat my previous question. Can you provide me with a reference to the "patent pending" application filling?


    Regards,

    William Louth
    CTO, JInspired

    "J2EE tuning, testing, tracing and monitoring with Insight"
    http://www.jinspired.com
  12. Here we go again.

    Simplified interface for the users doesn't mean that we don't have a solid understanding of the internals. This is one of the strong points of AutoPilot that we are accounting for all the complexities but still are able to provide our users an extremely easy to use product. In the interest of our viewers we had a similar conversation in Sept 2005 on dev2dev and here is the link to that.

    http://dev2dev.bea.com/pub/a/2005/09/performance_testing.html

    Performance Tuning of Application Servers & JVM is a well documented and proven process. Matt's article (link above) explains it well. You can also read more about it at the following BEA Document.

    http://edocs.bea.com/wls/docs81/perform/index.html

    At the same time for the last three days you have been spending time on our site that proves our products value right there :).

    We have proven it to our customers that 99.9 % figure is accurate. For you, it is something to figure out how we do it.

    AutoPilot is no silver bullet for the faulty code or bad architecture. What we do is get you the most out of whatever code you got rather than waiting and hoping to get that perfect code one day.

    From our experience most of the time (80 to 90%)is actually spent in narrowing down the problem to a particular area eg. jsp, ejb etc... Once problem is identified fixing it in most case is an easier task (10 -20 %). Quick and easy problem detection without requiring duplicating the problem in a test environment is what we have addressed with AutoPilot. AutoPilot's automatic root cause analysis capabilities identify the problematic area/areas and generates root cause report without any manual intervention at all.

    In terms of patents why don't you let USPTO do their job. We are free to apply a patent for whatever we believe is invented by us and it is up to Patent office do decide on the claim.
  13. "AutoPilot is no silver bullet for the faulty code or bad architecture. What we do is get you the most out of whatever code you got rather than waiting and hoping to get that perfect code one day."

    This is a rather naive approach to application performance management and the actual real-world scalability and performance problems faced by many companies today. If you are really going to simplify the problem domain to such an extent then you might as well recommend the "buy hardware option" if would probably be just as effective.

    > "At the same time for the last three days you have been spending time on our site that proves our products value right there :)."

    I see offerings like this, "99.9%", as potentially damaging to the performance and test management industry because in my past experience they over promise and under deliver when taken outside a contrived lab environment context.

    The big mistake made by most performance management vendors who rely solely on high level JMX metric collection is that they assume that production J2EE applications have been throughly tested under current (and changing) workloads and application usage patterns, and that any performance problems are caused by a small configuration change rather than a software and environment update.

    I have no problem with solutions that offer to some degree immediate and short terms performance fixes. It is always good to have a system configuration tune up but this is not a replacement for an application performance test and managment solution.
     
    In my opinion it is the container vendors, application developers, application architects and database/system administrators who have the knowledge to define the best practices. The real long term solution is to offer the ability to define (extensible) behavioral inspections which can be applied online and offline against a comprehensive execution model.

    If people are interested in understanding the problems and solutions offered by an enterprise performance test and management offering then they should also check the following performance articles.
    http://www.jinspired.com/products/jxinsight/insights.html

    > "We have proven it to our customers that 99.9 % figure is accurate."

    Surely before anyone becomes a customer they would like to know how a vendor could even begin to provide any sort of assurance - we could all start shouting 99.9999999% on our website but that would be unethical. Most vendors hope that their solution is effective in the majority of cases but stating an actual number implies that you have built test software that could verify this to be true for a large software application such as WebLogic 8.x and all the many applications running on it today.


    > "In terms of patents why don't you let USPTO do their job"

    We all have seen small startup companies talk about "patents pending" but in fact there was never an actual application filing. Apparently the reason it is placed on the website is to infer some kind of innovation, software magic, and worth.


    William Louth
    JXInsight Product Architect
    JInspired

    "J*EE tuning, testing, tracing and monitoring with Insight"
    http://www.jinspired.com
  14. We are putting our money where our mouth is. Here is how confident we are about the accuracy of the data published on our web site in the following statement.

    "AutoPilot Detector Auto-detects WebLogic Server failure conditions with 99.9 % accuracy."

    We are extending this offer to our customers. Work with us and let us show you how accurate & powerful our Auto Detection capabilities are. If the numbers don't add up to the claim you get our software license for free (we will be doing a press release with the details on this offer shortly).

    Go ahead download Arcturus AutoPilot WL today and try it for yourself.

    http://support.arcturustech.com/downloadpage.do

    Deepak Batra
    Founder & CEO
    Arcturus Technologies, Inc.

    Arcturus AutoPilot WL : The Smarter Way To Manage WebLogic
    Learn More: http://www.arcturustech.com/