An engineer is configuring the monitoring for a new application server. During the day, the CPU on the server is baselined at approximately 30% utilization. At midnight, a batch job is scheduled to…

Setting a different threshold during the batch job's scheduled time window creates a targeted alert exception that accounts for the known, expected CPU spike without altering sensitivity during normal operating hours. This approach eliminates message fatigue by addressing the specific, predictable cause of false positives while preserving full alerting capability for unexpected CPU issues at all other times. It is the most precise solution because it scopes the change to the exact period and condition responsible for the false alerts.

Increasing the alert time threshold to 65 minutes would cause administrators to miss legitimate, sustained CPU problems that persist for 10 to 64 minutes during normal business hours.

Raising the utilization threshold globally to 80% would suppress alerts for significant CPU spikes during the day that genuinely require investigation, reducing overall monitoring effectiveness.

Manually disabling monitoring during the batch job creates a complete visibility gap, meaning legitimate failures or unexpected errors occurring at midnight would go entirely undetected.