Skip to main content
Version: v1.0.0

Trend & Anomaly

Trend Analysis​

The Trend Analysis feature visualizes how infrastructure costs change over time. It provides clear insights into spending behaviour, highlights recurring patterns, and identifies gradual increases or decreases in cost. Trends can be viewed at Regular, Daily or Weekly levels, allowing both short-term and long-term cost evaluation. This feature sets the foundation for Anomaly Detection and Forecasting, helping user to track performance before predictive or optimization analysis.

Key Highlights​

  • Supports Regular, Daily and Weekly trend views for detailed or summarized seasonal insights.

  • Displays historical spending patterns.

  • Helps correlate cost fluctuations with deployment activities, workload changes, or scaling events.

  • Enables early detection of abnormal cost behaviour that may require optimization.

How to Use​

  • Navigate to the Trend & Anomaly page.

  • Select the desired Cluster.

  • Choose the time interval — Regular, Daily, or Weekly.

  • Review the generated trend chart to understand cost and usage variations over time.

🎯 Minimum Requirement: To generate Daily Trends, a cluster must be at least 2 days old (i.e., have 2 days of metric history). Weekly Trends can be generated for clusters with at least 1 week of available data.

Anomaly Detection​

The Anomaly Detection feature automatically identifies unusual cost fluctuations within existing data. It works in conjunction with the Trend Analysis view, displaying anomalies directly on the same graph to help quickly spot irregular behaviour without switching pages. Anomalies are visually represented on the trend line as colored dots indicating whether the detected cost is higher or lower than expected.

Key Highlights​

  • Detects sudden spikes or drops in cost compared to historical patterns.

  • Each anomaly dot represents a data point where the cost is higher/lower than expected.

  • Helps pinpoint abnormal activity caused by deployments, scaling events, or resource misconfigurations.

  • Enables faster investigation and timely optimization actions.

How to Use​

  • Navigate to the Trend & Anomaly page.

  • Select the desired Cluster and choose the appropriate time interval (Keep seasonality as Regular).

  • Observe the graph :

    • Coloured dots indicate anomalies where cost is higher/lower than expected.

  • Hover over any dot to view anomaly details, including the date, actual vs. expected cost.

  • Correlate anomalies with deployment or scaling activities to identify potential root causes.

🎯 Tip: Anomalies are identified based on cluster's available metric history. The longer the data, the more accurate and stable the anomaly detection results.