Rules

Anomaly Formula: Trend-Deviation

What it does:
This formula uses a linear regression to model the overall trend of your usage. It checks whether the latest value deviates too much from what the trend predicts—making it ideal for detecting subtle or extreme changes that defy the usual direction of the data.

When to use it:
Use this when you want to track changes against the expected trend, especially when usage typically follows a predictable upward or downward slope.

Examples

✅ Example 1: Alert Triggered (Decreased)

• Anomaly Formula: TrendDeviation
• Change Type: Decreased
• Threshold: 0.1
• History: 58, 84, 24, 96, 666, 14, 25, 23, 125, 39, 36, 48, 282, 10, 100, 25, 25, 28, 29, 215
• Explanation:
  • Based on trend, last value should be around 68.
  • Actual value is 28 → ~58% drop.
  • ✅ Alert triggered due to significant negative deviation.

✅ Example 2: Alert Triggered (Increased)

• Anomaly Formula: TrendDeviation
• Change Type: Increased
• Threshold: 0.2
• History: 20, 21, 22, 22, 23, 24, 24, 25, 26, 60
• Explanation:
  • Slow trend upward; final value jumps to 60.
  • Deviation from predicted is ~+122%.
  • ✅ Alert triggered for unexpected spike.

❌ Example 3: Normal Case (Any)

• Anomaly Formula: TrendDeviation
• Change Type: Any
• Threshold: 0.25
• History: 20, 23, 25, 28, 30, 33, 34, 35, 36, 38
• Explanation:
  • Data is rising gradually, last value continues trend.
  • Deviation is small, ~+4%.
  • ❌ No alert triggered.