Drinking our own Analytics Champagne

Here at HP Software we have a Demo Solutions Group (DSG) that’s in charge of creating demo systems that support presales and sales teams when demonstrating the products as part of the sales cycle. These live demo systems provide a real-life experience of the products, and are created to give the customer live examples of the value propositions of HP Software products.

Because our products need to run on a live software application, we created the “Advantage Online Banking” (AOB) application that supports the fake financial institution - “Advantage Inc.”. If any of our readers are current HP Software customers – you’ve doubtless seen this application during one or more of our demos.

 

While working with the AOB application, we found that it suffered from some sporadic problems, every now and then.  The application performance would become erratic and would subsequently crash.  The crashes seemed to occur after several days, but there were no assurances. This caused productivity issues for the Demo Solution Group, as well as for the sales and pre-sales teams. It meant the sales and pre-sales teams had to keep track of how long the application was up, and the Demo Solution Group had to assist with restarts and setups of the environment every now and then. This was a source of constant hassle, especially before presenting to key accounts.

 

Over the years we tried to investigate the issue – we looked at the DB, we analyzed the back-end and front-end servers, we mined the log files, and we consulted with the application experts. Unfortunately none of these avenues yielded any results.

Enter Operations Analytics 2.1.

Now that we had the new version of Operations Analytics we had to create some new demo scenarios to clearly show the value of the new features – especially Log Analytics.

Log Analytics enables automatic location of the needle in the log messages haystack. It uses pattern matching and machine learning algorithms, in addition to crowd wisdom, to narrow down the millions of logs to the few messages that really matter when you’re trying to find the root cause of a problem. The exciting thing about this feature is that you no longer have to guess which term to search for, create complex regular expressions as part of your search, or engage your most sought after system experts to investigate. Log Analytics power is finding the information on its own without this time consuming hassle.

So here we were trying to create a problem in our demo application, a problem that also created some relevant log messages. This was all done so that Log Analytics could automatically find these significant messages, and we would be able to demonstrate the power of this new feature to our customers. So we set up the system – the Operation Analytics product, and the demo application “Advantage Online Banking” it would run on. We were planning to look for an interesting use case the next week.

When we came back from our weekend planning to look for the right use case for the new Log Analytics feature, we found we didn’t need to look any more.

 

We found that during the weekend the AOB application was experiencing those same problems we’d been experiencing for about three years.

 

However what amazed us was that in addition to the problem Operation Analytics had detected, the Log Analytics capability, had pinpointed the root cause of the problem. Log analytics had analyzed all the log messages, and automatically detected an error that caused a transaction rollback and subsequently the performance degradation we’d been seeing.

fig1.jpg

Figure 1: OpsA detect the perfomance issue on AOB application

 

fig2.jpg

Figure 2: Log analytic identify 6 significant log messages out of 2,863 during 1 hr time frame

 

fig3.jpg

Figure 3: log message which pinpoint the problem in the app

 

This was the information we’d been looking for, and even though we used log search capabilities in the past, we’d not been able to locate the message as we obviously had no idea precisely what we were looking for. Log analytics had done the work for us.

The search was finally over, we found the problem on our AOB and even better, we found the right (and real) use case which exemplified the value of OpsA and specifically the log analytics part.

But because this example is so good, we decided to keep the AOB app the way it is, perfectly broken….

So now we’re drinking our own champagne, and it’s time for you to try out Operations Analytics 2.1, and some of its new exciting features.

Come and drink champagne with us. We promise you won’t regret it!

To learn more about HP Operations Analytics visit us at http://www.hp.com/go/opsanalytics, and check out how Operations Analytics can help you by checking out our Digital Demo Book at

http://www.hpdemodigitalbook.com/opsa

 

This Blog was written together with Yair Horowitz Chief Functional Architect – Operations Big Data Analytics

 

Comments
HP Expert | ‎07-03-2014 02:09 PM

Nice to see how it really works. Good that you left it broken for the use case :-)

Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the Community Guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
Showing results for 
Search instead for 
Do you mean 
About the Author


Follow Us
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation