That for parameter in your alert definition is your best friend for preventing “flapping” alerts (alerts that bounce between firing and resolving), which can quickly lead to alert fatigue – and nobody wants that! groups : - name : Blackbox Alerts Stream interval : 1m type : prometheus rules : - alert : HTTP Blackbox Failing Stream expr : 'http_response_result_code{check_type="normal"} > 0' for : 2m annotations : summary : '{{ $labels.server }} is not healthy' description : '{{ $labels.server }} is not responding properly from {{ $labels.host }}' - name : Blackbox SLO interval : 5m type : prometheus rules : - record : http:avg:latency:5m expr : 'avg(http_response_response_time[5m]) by (server)' Monitor Your Alerts: Once an alert condition is met and the for duration passes, the alert will…