Here’s an introduction to the newest features in HP Service Health Analyzer making predictive analyticsan even more powerful capability and easier to use. You can find more information at www.hp.com/go/sha
Article developed by Udi Shagal - HP SHA Product Manager and Ian Bromehead Sr Product Mkt Manager of HP Software.
Anomaly highlights page - overview of the business impact and probable causes including links to relevant BSM reports (based on the actual monitoring deployed).
Note how SHA compiles this business impact view, showing number of users this anomaly is impacting (as measured by RUM), the SLA status and the locations impacted (as measured by BPM).
2. New application screen – Open Anomalies This allows easy access to all open anomalies, with their severity, impact, suspected root cause…
3. Simplified deployment of SHA:
- SHA components are installed by default with the BSM 9.2x platform.
- SHA no longer needs a separate Analytics Server (besides some documented use cases)
4. Improved anomaly isolation - better qualification of probable cause (e.g. the CI with the most
abnormal metrics) based on the correlations as can be seen in the correlation tab below.
5. Directional baselines – reduces false positives by ignoring baseline violations where the violation direction is less likely to indicate a problem. For example, an unusually short response time might not indicate a problem.
6. Baseline sensitivity per domain – reduces false positives by allowing different sensitivity settings for
different domains instead of a generic sensitivity setting for all domains. For example, some of our customers
noticed that their system metrics are much noisier than their EUM metrics. With this setting you no longer need to
7. Better detection of anomaly closure – algorithm improvement leading to less open anomalies to be investigated.
8. Anomaly severity control - users can now define which metrics should be considered as business critical. If those metrics breach their baseline, the severity for the detected anomaly would be set to Critical.
9. Better support for shared CIs – added a setting which allows users to define the desired anomaly scope in the case of shared infrastructure CIs. For example, if the same database server supports multiple business applications and abnormal metrics are detected in those applications, SHA can ignore the shared database relationship and create smaller and more focused anomalies per application vs. one big anomaly for all applications. We recently tested this enhancement on a large scale production environment and it had a great positive impact on the
usability and performance of SHA.
10. Extended similarity analysis – a larger set of past anomalies will be considered for similarity analysis
(vs. the last 500 anomalies which exists today). For large scale environments, this results in a better
chance of anomaly similarity matching.
11. Investigation user interface - Improved filtering capability, allowing the user to filter the topology view model only to the CIs with abnormal metrics (see screenshot) vs. seeing the entire application model. For large models the filter is applied automatically to help improve the UI performance and usability.
12. Support for 3rd party metrics via BSM Connector. In previous versions we had 2 approaches for 3rd party integrations – EMS and Integration Adapter. In version BSM 9.20 we merged the best of these 2 solutions and created a more efficient solution, attractive interface and a common development environment. Some of our customers have already exploited these new capabilities to integrate various 3rd party data with little effort
(e.g. metrics from Informix DB, BMC Coradiant, etc) and feed those metrics into the analytics done by SHA.