Anomaly Detection

Solution Capabilities & Features

Overview

You receive a barrage of alerts every day and over 90% of these alerts you ignore reflexively. This is despite the risk of negatively impacting team productivity and MTTR metric by prematurely acting on the wrong alerts and dismissing consequential warnings. This results in Alert fatigue. The root cause of alert fatigue is reliance on a hodgepodge of unintelligent, legacy tools used to monitor a growing landscape of systems and solution stack layers. AI-powered Anomaly Detection can eliminate your alert fatigue pain while addressing key business drivers.

Anomaly Detection Business Drivers

Full-landscape
visibility

Automated process
 improvement

Manage-by-exception
efficiency

Key AIOps Capabilities & Features

Behavior Profiling (AutoProfile™)
  • Develops behavior profiles based on consolidated full-stack data collection, statistical modeling, historical data, and text analysis
  • Leverages numerical unstructured and contextual data
  • Generates profiles at multiple levels of granularity, such as landscape, systems, and components (e.g. KPIs/performance metrics for each unique app, server, and DB component)
  • Recognizes and adapts to normal changes in system behavior over time (e.g. different hours of the day, different days of the week)
Dynamic Thresholds
  • Derives thresholds based on continuous monitoring and consideration of expected normal behavior
  • Adjusts thresholds automatically rather than using static thresholds which take a long time to configure; can be inaccurate if SAP default thresholds are used; and need to be adjusted manually
  • Supports ability to set policy-based manual overrides of dynamic thresholds on the fly
AI-Guided Scoring
  • Prioritizes anomalies for further evaluation based on a holistic view
  • Scores anomalies based on incident frequency and magnitude (i.e. difference between medium of data points and threshold value)
  • Ensures through Adaptive Clustering™ that only significant clusters of incidents within a time window interval contribute to a higher score and not every isolated threshold breach

Next Steps

1.  Check out the Anomaly Detection Solution eBrief

2. Contact us to initiate a needs discovery dialogue

3. Request a demo

Back to top