IBM is a leader in providing predictive analytics solutions for your IT infrastructure. These capabilities are integrated into all Tivoli monitoring solutions. Predictive analytics capabilities increase the availability of your IT infrastructure by proactively providing information regarding potential infrastructure issues before your business or IT environment is negatively affected.
Typical issues a customer faces when implementing a monitoring solution include:
- What to measure, i.e. what thresholds to set in order to be notified of a performance issue?
- Ensuring that the notification happens in a timely manner
- Understanding how a performance issue impacts a service
When customers first deploy a performance management solution, they go about defining thresholds on KPIs to enable notifications when performance is starting to impact their customer’s services. Not only is manual threshold management very time consuming, it is error prone as well. If the threshold is set too low, operations is faced with many “false positives”. However, if the threshold is set at the peak to avoid these alarms, then many performance issues will not be detected until too late.
One of the approaches that IBM Tivoli takes to address this situation is by providing dynamic baselining capabilities. With this approach, the system learns what the normal behavior is for this resource’s performance. This includes the seasonality associated with the hourly and daily deviations of use. With this approach, the system learns the normal behavior and adjusts this over time. When the performance starts to deviate from the norm, a notification will be raised.
Dynamic thresholding complements another Tivoli predictive analytics capability, forward trending. Forward trending allows a customer to visualize a performance metric’s forecasted values and take appropriate action before a threshold is even reached.
Benefits
- Advance warning of service impact, deterioration or outage
- Realistic service baselines
- Avoidance of expensive and time-consuming false alerts
- Detection of service impacts that are not identified by fixed thresholds alone
- Swifter diagnosis of certain events and patterns
- Identification of the underlying root cause to implement fixes
