Predicts performance degradation of critical services in real time, enabling customers take preventive action.
Real Time Prediction Accuracy is our hallmark
Our predictive technology looks for fingerprints in the data that were indicative of previous problems and alert when there is a re-emergence of the same fingerprint. Our approach inherently increases confidence and accuracy of predictability relative to competitive approach as the same data pattern indicative of a previous occurrence of problem is very likely to indicate occurrence of a problem again.
Quantification of Possible Impact
Our dashboard displays changing probability of risk of interruption to Business Services in real-time, allowing customers to decide on appropriate course of action. So, as an example we might indicate a probability of 30% twelve hours in advance and as time progresses probability might increase to 90% in 2 hours. System can additionally be configured to alert only when certain time and confidence levels are reached. This minimizes predictive false alerts drastically and customers will be able to interpret the predictive alert based on their specific environment.
Delence’s patent pending Alert Rules Generation Engine utilizes multi-dimensional analysis and pattern recognition engine to generate rules that would indicate likelihood of interruption to a business service. These rules can be updated at regular intervals (weekly or monthly) to learn from the most recent slow-downs and outages. This Engine will typically reside on Delence’s server farm, churning updated alert rules as soon as more recent historical data is available for processing. Updated rules can be set to automatically update the real time light weight query engine, if necessary.
A light weight real-time query engine, resident in client’s infrastructure continuously queries client’s monitoring data store and log files for patterns that match the alert rules. A match will indicate a likelihood of interruption to a business service. When a match is identified appropriate personnel are alerted with a probability of occurrence, an estimated time of interruption to a business service and possible solutions to predicted issue.