Fraud Rules Configuration

Configure fraud detection rules and thresholds

Total Rules

8

Active Rules

8

Real-Time Rules

6

Batch Rules

2

Detection Rules
Rule Name Type Severity Status Thresholds Action Detections (30d) Actions
Missing Impression
Detects clicks without matching impression 		real time critical
None block 4
IP Frequency
Detects excessive impressions from single IP 		real time high
Max Impressions: 60
Time Window Minutes: 1
flag 8
Click Timing
Detects clicks too fast or too slow after impression 		real time high
Min Seconds: 3
Max Seconds: 900
flag 1
User Agent Validation
Detects bots and malformed user agents 		real time medium
None flag 9
Geographic Anomaly
Detects impossible geographic patterns 		real time medium
None flag 3
Referrer Validation
Validates referrer authenticity 		real time low
None log 4
Pattern Detection
Detects statistically anomalous patterns 		batch high
CTR Threshold: 0.15
CVR Threshold: 0.10
flag 7
Session Analysis
Detects excessive activity per session 		batch medium
Max Impressions Per Session: 80
flag 3
Rule Types & Actions
Rule Types
  • Real-Time: Executed during ad serving/tracking requests. Fast evaluation (<5ms). Blocks requests immediately.
  • Batch: Executed daily via cron job. Analyzes aggregate patterns. Flags suspicious activity for review.
Actions
  • Log: Record detection in database (no blocking)
  • Flag: Mark for manual review in dashboard
  • Block: Prevent ad serving for this request
  • Blacklist: Automatically add IP to blacklist
Available Rules
IP Frequency: Detects excessive impressions from single IP
Click Timing: Validates click timing after impression
Missing Impression: Blocks clicks without valid impression
User Agent Validation: Detects bots and malformed UAs
Geographic Anomaly: Validates geo consistency
Referrer Validation: Checks referrer authenticity
Session Analysis: Detects excessive activity per session
Pattern Detection: Finds statistically anomalous CTR/CVR