Alert fatigue is reduced attention or responsiveness caused by excessive, repetitive, or low-value alerts. In monitoring systems, overly sensitive thresholds create false positives that can eventually make operators ignore genuine problems. section 15.4.1
Reduce alert fatigue by treating alert configuration as an ongoing operational practice:
- Set meaningful thresholds based on the metric’s past behavior rather than choosing excessively sensitive limits. section 15.4.1
- Adjust thresholds when false positives appear, then regularly review and update the configuration. section 15.4.1
- Smooth metrics and add contextual information so alerts represent real problems and support faster diagnosis under pressure. section 15.6
A noisy monitoring threshold
A Kafka monitoring system configured with an excessively sensitive threshold repeatedly reports harmless metric fluctuations. Operators begin ignoring the notifications, so a later alert for a real processing delay may be overlooked. Raising the threshold, smoothing the metric, and reviewing the rule regularly makes notifications more actionable. section 15.4.1 section 15.6
