How to Create CPU Usage Alerts in ControlUp That Exclude Restart and Boot Spikes

A common challenge in monitoring CPU usage is avoiding false alerts during system restarts or boot periods, when CPU spikes are normal but typically short-lived. One effective method to address this in ControlUp is to configure an alert rule targeting the device_status data index with specific conditions to filter out these transient spikes.

To create a CPU alert that excludes restart or boot-phase spikes, set up a static threshold alert on Devices with a condition such as cpuload > 80 (adjust the threshold based on your environment). The key element is applying a time window and hit count that accommodates short bursts without triggering alerts. For example, use a time window of 600 seconds (10 minutes) and require 5 hits within that period before the alert fires. Additionally, a retrigger delay of 3600 seconds (1 hour) helps reduce repeated alerts for sustained high CPU usage.

This configuration works because CPU usage during boot or restart typically spikes briefly and does not persist long enough to meet the hit criteria within the time window, thereby filtering out most false positives without complex logic. For situations where CPU remains elevated for longer periods during boot—due to updates, login scripts, or heavy startup applications—it is recommended to implement an uptime check or guard. This further refines alerting by ensuring the system has been running beyond a certain uptime threshold before alerts activate.

These steps provide a practical approach to balancing sensitivity and noise in CPU alerting, improving monitoring accuracy in environments where restarts cause predictable, short-term CPU spikes. For more detailed guidance on creating and tuning alert rules, consult the ControlUp documentation at https://docs.controlup.com/alert-rules and consider exploring uptime monitoring techniques available in ControlUp’s device health metrics.

Read the entire ‘How to Create CPU Usage Alerts in ControlUp That Exclude Restart and Boot Spikes’ thread below:

Good day
Anyone perhaps build a alert that will detect High CPU but it will not alert during a restart or a boot?

hey Eugene how’s u. You can try this

CPU alert on `device_status`:
• Events → Alert Rules → Add Alert
• Product: Devices |Method: Static Threshold
• Data index: device_status
• Condition:`cpuload > 80`(tune as needed)
• Time window:`600`seconds (10 min)
• Hits:`5`
• Retrigger delay:`3600`seconds
Boot/restart CPU spikes are usually short. Requiring 5 hits inside 10 minutes filters most false positives without extra logic.
If boot CPU stays high longer (updates, login scripts, heavy startup apps), add an uptime guard.

Continue reading and comment on the thread ‘How to Create CPU Usage Alerts in ControlUp That Exclude Restart and Boot Spikes’.  Not a member? Join Here!