The importance of J2EE applications and their relevance cannot be over emphasized. More and more companies are increasingly relying on them both for their frontend and backend processing. While the software community has made a lot of positive strides in providing frameworks (like standardized J2EE containers) to develop and host these applications, the solutions that are being deployed for managing and monitoring these applications are still a few steps behind. This mismatch has caused unwarranted downtime, slow, painful and expensive turnaround time and in almost all cases breaches of Service Level Agreements.

The existing tools or the conventional monitoring techniques fall short and don't reliably and easily help pinpoint and isolate issues. In absence of good monitoring support, the IT teams often resort to the guess game and invariably use trial and error approaches which is time-consuming and unreliable.

deterMyne's insideApps uses recent advances made in JVM in the field of monitoring and diagnosis to give an inside out view of the system components. Unlike traditional profiler, which runs along with the application, insideApps uses "Tag & Follow" technique and deploys Instrumentation Layer [IL] in the production nodes. These ILs collect real-time information and send it to central console. The central console is responsible for correlating the data from various IL's and provides more flexible management and monitoring features. As the resource intensive operation of correlation happens in a separate machine, the performance of the actual application is not affected.

The key features provided by insideApps are:

• End-to-End Business Transaction Monitoring including hops between HTTP server, multiple J2EE containers and databases.
• Service Level Agreement [SLA]
• Baseline support
• Helps in overcoming the traditional "LAST MILE" problem.
• Root Cause Analysis Engine [RCA]
• Quick turnaround in resolving the issues
• High valued transaction monitoring - Monitor transaction
• specific to a customer or an operation
• Inbuilt alert engine
• Discover the application and topology dynamically based on request routes
• Applications Dashboard

For more information and realtime demo, please visit determyne.com.

Some questions unanswered by this article or web site:
* what is the impact of monitoring, in term of CPU & memory performance?

• how much does it cost?
• how easy is it to install/use?
• where can I download a demonstration version? :-)

As for a light-weight alternative for monitoring the Web tier, have a look at MessAdmin, my very own open-source tool (shameless plug :)).

what is the impact of monitoring, in term of CPU & memory performance?

The resource utilization is less than 2% in all our test environments. As the complex process of analyzing the data from different instrumentation layers happens in a different machine, the performance of the application wont get affected. Please refer the white papers published in our website for more information. We are planning to publish a data sheet in this regard.

how much does it cost?

We are an exciting startup aiming to provide the best and most affordable performance monitoring solution for applications. We are in the middle of developing very close relationship with a handful of customers. We hope to work very closely to ensure that their needs are met as well as it gives us an opportunity to incorporate feedbacks received and tailor the products accordingly. For our customers, we will allow them to determine the price based on value propositon they see having used it. If you are not satisfied, no need to pay a dime. Please send an email to info at determyne dot com for more information.

how easy is it to install/use?

Installation and use is very simple. Please take a look at our realtime demo and try yourself.

where can I download a demonstration version? :-)

We are planning to publish the evaluation version of our product soon. Please check determyne.com and register yourself to receive the details.

Hi "Neo",
How is it different from JXInsight?

Thanks,
Nikita Ivanov.
GridGain - Grid Computing Made Simple

Hi "Neo",
How is it different from JXInsight?

Thanks,
Nikita Ivanov.
GridGain - Grid Computing Made Simple

We believe that success of a monitoring product depends on a variety of factors -

1. Richness of metrics and level of corelation.
2. Root cause analysis capabilities.
3. Implementation - small footprint and low overheads.
4. Intuitive design of UI
5. Actual cost of ownership and support

insideApps has been introduced keeping all these factors in mind. Apart from the above, a few of the following features are big differentiators compared to other products -

1. Runtime topology - this is a dynamic topology derived by following the requests using our intelligent "tag and follow" technology. insideApps allows you to view the whole application topology as well as at individual transaction level.
2. Rich Database metrics - insideApps goes inside the database to give contextual information about each and every request. It has details about wait conditions and exec plan explained and tries to find out what in database might be causing bottlenecks.
3. Call trace - insideApps has method level tracing support for not only within one container but across all the containers involved in a particular request. It gives time spent on each methods.
4. Other intuitive feature like saving and comparing against baselines to find out what has changed.
5. Support for Apache HTTP metrics - insideApps could tell you how much time was spent in the HTTP server also.

Thanks,
Neo
www.deterMyne.com

Hi Neo,

I do not think you have actually answered the question. All of the factors you have mentioned could be attributed to design of any number of other products though the quality of the design and implementation does vary. In fact to be honest I think it would be extremely hard to out perform JXInsight in any of these areas apart from giving away the product free and offering free training and support but I suspect this is not your business plan.

Maybe I am wrong. I have not seen your product as I am not going to fill out a form with my contact details just to see a screenshot of the product or to see more than a few lines of bullets points on key features.

For your information JXInsight can provide contextual trace across tiers. We have demonstrated on articles and blog entries this WebLogic RMI, JBoss Remoting, CORBA, GridGain, Coherence, and a few other more specialized platforms.

I think you will find that when it comes to Java runtime diagnostics and extensibility our product is (currently) unbeatable. That was our focus up to JXInsight 5.5. Which is the focus of this website.

No Java related metric analysis has been targeted for JXInsight 6.0 "Metrica".

http://blog.jinspired.com/?p=209

We already provide a very large number of JVM and software execution related metrics in JXInsight 5.5 and our correlation is not just visually.

http://blog.jinspired.com/?cat=18

You do mention diagnostics but I am not sure if this is a very primitive form of metric analysis or do you offer the ability to inspect runtime state of inflight requests as well as system/component state.

http://blog.jinspired.com/?cat=19

In terms of overhead in production up to now we have won every shoot off with our Probes resource metering solution which offers a powerful set of metering strategies that can be combined. Note: We recommend probes over tracing in production because if you do not know how your software executes when it comes to a production you are already in serious trouble.

http://blog.jinspired.com/?cat=21

You do mention database metrics but I would prefer true resource transaction analysis over database metrics if I had to choose but then again I am focused on the software & system execution models and not just system monitoring because I spent more time architecting and building systems and less time staring at big green and red circles.

Without a thorough understanding of a software's execution patterns root cause analysis via metrics is extremely inefficient involving a large amount of guess work.

http://blog.jinspired.com/?p=39

William Louth
JXInsight Product Architect
CTO, JINSPIRED

sounds interesting. Can the product also trace requests between multiple J2EE node and corelate with database activity? I am asking this because our applications typically have two levels of J2EE containers, first one typically calls EJB in the second J2EE container and then the database. For redundancy we have multiple of these nodes and it is a practical nightmare trying to trace problems - and corelating the events. The website seems to indicate that you have this covered.
Also do you support .net applications also?

sounds interesting. Can the product also trace requests between multiple J2EE node and corelate with database activity? I am asking this because our applications typically have two levels of J2EE containers, first one typically calls EJB in the second J2EE container and then the database. For redundancy we have multiple of these nodes and it is a practical nightmare trying to trace problems - and corelating the events. The website seems to indicate that you have this covered.

Yes. We support end to end business level transactions and show breakdown across different nodes. We also show database queries and reasons for database performance problems. For e.g., if a HTTP request resulted in a database query and that query gets blocked because of some lock condition inside the database, insideApps would show that detail. Please see demo and use cases doc for more information.

Also do you support .net applications also?

For .net, it is in our radar screen but we dont have that support yet.

Does it co-relate every single http request performance with the database performance?

If yes what sort of metrics are provided? Does this help me in optimizing the sql queries?

We in our application use Hibernate, does this monitoring tool provide breakup in terms of packages or frameworks where it spends most time?

Is it possible to find for a single http request how much time it spends in Hibernate?