I wanted to show you how easy it is now to get the best IT service analytics capabilities in the industry. In fact, it’s hard to imagine how much simpler it could be to deploy the new HP Service Health Analyzer 9.23 (SHA). You can get started by downloading SHA here.
Existing users of the HP Business Service Management (BSM) platform will need to spend just a couple of minutes on configuration. That’s it. You do not require any extra hardware.
So let me take you through what kinds of metrics you will start to be able to track after just half a day. I also have a few tips to share for using SHA to boost your analytics capabilities and drive more value in less time.
SHA requires almost no configuration and zero maintenance because it uses your existing BSM platform. All you need to do is define the Analytics database in which SHA will store its raw data and baseline.
Next, you will define the business services or applications you would like to analyze. This is a simple drag and drop process.
And you are all set! Now, sit back and let SHA work its magic.
What you will know after 30 minutes
Within the first half hour after you finish configuring SHA you will be able to start see all the CIs and their metrics that SHA monitors.
SHA will present a holistic view of your application or service, generating its topology together with a time machine that displays all changes (both planned and discovered) as well as events and incidents that have occurred (Figure 1).
Fig. 1: SHA topology view
What you will know after two hours
Within a few hours, you will be able to view all the metrics that SHA monitors and begin to perform correlation analysis both by using SHA advanced correlation tools or by using SHA graphs to overlay metrics for comparison (Figure 2).
Fig.2: SHA correlations
What you will know within the first 12 hours
After less than half a day, SHA will automatically calculate and establish intial baselines for each of your metrics, so you can start enjoying the full capabilites of SHA.
It will start detecting abnormal behavior in your IT environment, estimate the potential business impact and apply several algorithms to isolate and pinpoint root causes, presented in a powerful UI.
How SHA improves over time
As with any analytic tool, the more data you input, the more it improves. We have designed SHA to start offering you value after less than 12 hours, but as time passes, its accuracy increases. Here are a couple of examples:
SHA applies a unique algorithm to identify seasonality trends in your data, which allows it to set different thresholds based on the times of day or week, and predict how your applications ought to be performing
Through trend detection, SHA predicts expected increases or decreases in your metrics, and indicate whether such a trend exists
Tip #1: Increasing seasonality and trend calculations
SHA was built to work on top of “big data” (millions of metrics) for an entire IT environment with 20,000 or more nodes, so by default it is scheduled to calculate seasonality and trends once a month. But when your first activate it, you may monitoring only a handful of services and calculating a few hundred-thousand metrics on about 2000 nodes.
If that’s the case, we recommend that you speed up the seasonality and trend calculations by changing its frequency from 1 month to 7 days.
To refresh seasonality and trends every week:
Browse to: http://<enter your BSM gateway machine>>:8080/jmx-console/HtmlAdaptor?action=inspectMBean&name=Foundations%3Aservice%3DInfrastructure+Settings+Manager
Go to setGlobalSettingValue and apply the following values:
contextName = baselineSettings
newValue = 7
And then apply the same values for a trend:
contextName = baselineSettings
Tip #2: Pattern recognition
Last but not least, SHA applies pattern recognition for all the anomlies it detects. It is strongly recommended that for each anomaly you investigate, spend an additonal minute and consider marking it as a pattern if you think that there is a chance that this kind of anomaly with the same “DNA” will repeat in the future. This will help you dramatically speed up your investigation a similar anomaly is detected in the future.
You can also mark some anomalies as “noise”, to automatically reduce its severity rating should SHA recognize a similar anomlay again. This will help you train and fine-tune SHA specifically to your IT environment.