What happens if my thresholds are too sensitive (false positive alerts)?
Use Statistics to analyze actual patterns and adjust thresholds. False positive symptoms: Alerts firing during legitimate low-volume periods (weekends, holidays, after-hours), minor variances trigger alerts (order volume fluctuates 745-855/hour, threshold Min 800 triggers intermittent alerts).
Tuning process:
- Collect 2-4 weeks of Statistics: Review average message counts, standard deviation, Min/Max observed values, day-of-week patterns
- Identify false positives: Check alert history, correlate alerts to business events (did order volume actually drop, or normal Saturday low volume?)
- Adjust thresholds: Set Min threshold to 80% of average (allows 20% variance), set Max threshold to 150% of average (50% buffer for peaks). Example: Average 847 orders/hour → Min 678 (80%), Max 1,271 (150%)
- Refine business hours: If false positives occur weekends, change schedule from "7 days/week" to "Mon-Fri only", set separate weekend thresholds (lower Min)
- Add exclude dates: Import company holiday calendar (Thanksgiving, Christmas, New Year's), exclude from evaluation
- Use Warning vs. Error thresholds: Set Warning at 85% of average (early heads-up, low urgency), Error at 70% of average (critical data outage, high urgency). Reduces alert fatigue (fewer critical alerts)
Statistics-driven example: Manufacturing production monitoring showed 12 false positive alerts/week (every Saturday/Sunday = "Machine Idle >60 minutes" alerts during scheduled no-production days). Solution: Added exclude dates Sat-Sun, false positives dropped to zero, alert accuracy 100%.
Next Step
Need more help? Check the Troubleshooting Overview for all FAQs, or refer to the Configuration Guide to fine-tune your Non-Events monitoring setup.