IBM have designed and developed Predictive Insights from the ground-up and because of that it has been able to respond to customer demands and requirements for a real-time performance analysis tool. But why should your company consider using Predictive Insights? Here are 10 reasons why Predictive Insights can swiftly help you improve the quality of your enterprise monitoring, reduce incidents and save money.
1. Data agnostics
Predictive Insights can ingest any time-based metrics, i.e. data where values are sample over time. And so, for example, it can analyse and monitor metrics collected from applications, platforms, networks and SANs.
2. Leverage the data from your existing monitoring solution
Mediation Packs are available to integrate with many of the leading monitoring solutions, including, but not limited to Microsoft SCOM, Solarwinds, VMWare vCentre, Splunk, AppDynamic, Dynatrace, DC RIM and Aternity. The mediation tool can be used to integrate to other data sources.
Predictive Insights requires no monitoring configuration. It automatically learns the behaviour of metrics, how they change over time and their relationships. It automatically identifies the best algorithms to use for modelling the data. One of the principles behind Predictive Insights is “why ask a question when the computer can work it out”.
4. Identify relationships across technologies
Predictive Insights learns relationships across technology boundaries, for example, the relationship between metrics for a web service response time and database pool hit ratio. Alerts will be generated if there is abnormal behaviour within those learnt relationships.
5. Real-time data analysis
Predictive Insights analyses real-time data from the various data sources and will identify “anomalies” in the data, i.e. where the real-time data does not align with the expected, learnt behaviour.
6. No more threshold definitions
The complexity of threshold monitoring within an enterprise means that generic thresholds are often employed, thresholds that are inadequate for accurate monitoring. Predictive Insights automatically calculates the suitable thresholds for each metric on each resource. Those thresholds may be dynamic across the business day and week resulting in fewer false alerts yet ensuring problems are not missed due to thresholds being set too high.
7. Earlier alerts, fewer incidents
The combination of more appropriate thresholds and monitoring of metric relationships means that application and infrastructure problems are detected earlier than with traditional monitoring and with zero configuration effort from the ESM Team and the SMEs. This early warning enables problems to be resolved before they cause an incident.
8. Resolve incidents quicker, fewer alerts for the operations team
Metrics relating to the degradation of performance of an application or service, for example user response time, are the symptom of an incident but not the cause. Predictive Insights groups together alerts for related metrics enabling the context of such an alert to be understood, and the underlying cause to be quickly isolated.
9. Adaptive threshold monitoring across the business day and week
The self-learning feature of Predictive Insights will automatically define dynamic thresholds for metrics, where appropriate. These thresholds are based on the normal behaviour of the metric overtime for the given resource, and eliminate the inadequate monitoring associated with manually defined static thresholds.
10. Bi-directional Integration with Netcool/OMNIbus
Predictive Insights integration with Netcool/OMNIbus is two-way. Predictive Insights generates alerts in Netcool/OMNIbus when anomalies are detected. The alerts include hierarchical relationships. To assist with problem isolation the Operator can drill down to view the the underlying related child alerts or to the Predictive Insights service diagnostics chart.
How to Start your Predict Analytics Journey
How can you prove the worth of Predictive Insights without investing time and money in a protracted PoC? Its a straight forward exercise to analyse historic mertics to understand what problems would have been identified and why. This has been an revealing exercise for many customers.
If you are interested in investigating how Predictive Insights could help you then please email Orb Data at email@example.com.