Analytics to the rescue: How you can cut MTTR by half and save weeks of work

Demonstrating the potential of HP Service Health Analyzer (SHA) analytics to customers has always been a bit tricky. How can you prove that analytics software can help resolve and predict problems when that software first needs to learn what is “normal behavior”, and then wait for a “perfect storm” to come along?

 

One solution is to run a “what if” scenario for an incident that occurred in the past, and simulate what would have happened if the customer had had analytics in place. During the last few weeks, I’ve been part of a task force working on just such a Proof of Concept (PoC), and I think you’ll find the results to be really interesting.

 

You can download HP Service Health Analyzer here to experience it for yourself.

 

Designing the Proof of Concept

The challenges with a “what if” approach were two-fold:

 

  1. Some customers will save data from the time of an incident, but most will not save enough data from before the incident, which means SHA does not have enough data to learn what normal behavior is
  2. SHA is designed for real-time analysis; its capability to read historical data is limited

 

We were fortunate to work with a customer that, in fact, saved all its data. However, we still faced the problem of working on historical data with a product designed for real-time use cases.

 

As it turned out, the task was easier then we had expected. Here’s how we did it:

 

First, we extracted all the relevant data from the database and manipulated the time of each sample, shifting it by two months to simulate as if the problematic week was actually happening today. 

Then we used a 3rd party data collector to read the manipulated data and play it to Business Service Management (BSM), so that from BSM’s and SHA’s perspectives, the anomalies were occurring right now.

 

The results were a perfect demonstration of how SHA and Ops Analytics can help your IT. Not only was SHA able to automatically detect the problem (see Figure 1 below), but it successfully focused on the right root cause of the problem.

The customer estimate that using SHA tools he will be able to find the root cause of the problem at least twice as fast than using his existing tools.

 

SHA on SAP.png

Fig. 1: The first anomaly SHA detected, correlating all the relevant data into one place

 

SHA detect the problem without any configuration from the customer side, In fact we showed that SHA dynamic thresholds were equal and even better than the customer static thresholds, From the customer perspective it showed that he can reduce dramatically the on boarding of the his applications and just let SHA do the work for him.

 

Seeing the problem

Just as importantly, we were able to use our Ops Analytics to establish and clearly visualize the cause of the problem. In Figure 2, you can see the Ops Analytics report of the customer data showing the time of the incident and relevant metrics together with visualization.

 

OpsA on SAP.png

Fig. 2: Ops Analytics report presents the top utilized tables and processes

 

We used OpsA’s ability to show at any given time what the top problematic processes are, and point out the problems at hand. SHA provided the customer with a clear picture of what went wrong, coupled with analytics tools to root out the cause.

 

Getting results

In the end, the final report on the Proof of Concept confirmed three important results:

  1. The PoC was able to “bubble up” related metrics in a way that is usable for similar, yet not identical, issues
  2. Estimated 50% reduction in MTTR for SAP Performance Issues
  3. Estimated reduction of more than 10 weeks of work for each application on boarding

 

All in all, I think the PoC was quite successful, demonstrating not only the importance of saving all your data, but how analytics can help in a time of need.

 

Get more information

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
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