7. Tuning alert thresholds
Your LogicMonitor account comes with default thresholds for most devices, which means that alerts will be triggered right out of the box. You may need to adjust these thresholds to better fit your use case.
You can customize alert thresholds at the device level, the global DataSource level, or the device group level. Changes made at a narrower level will override any changes made at a broader level, with the global DataSource level being the broadest level and the DataSource instance level being the narrowest level.
Tuning thresholds at the global level
To change alert thresholds at the global DataSource level, select the DataSource in your device tree and then select edit global definition:
Remember you can set three thresholds for each datapoint in the DataSource definition—one for each of warning alerts, error alerts, and critical alerts. To change the alert thresholds for a datapoint, select the manage gear icon next to the datapoint name. Any changes made to the DataSource definition will apply to all instances of the DataSource in your account.
Tuning thresholds at the group level
To change alert thresholds for all DataSource instances in a particular device group, navigate to the Alert Tuning tab for that group and select manage threshold for a particular datapoint:
Tuning thresholds at the device level
Device level thresholds will override both group-level and global-level thresholds. You can set device-level thresholds using the Alert Tuning tab for a particular DataSource or instance depending on whether you are setting the threshold for a multi or single instance DataSource:
Tuning thresholds for single instance DataSources
To change the alert thresholds for one particular instance of a DataSource, instead of for all instances of the DataSource, navigate to the Alert Tuning tab for that DataSource instance and select the manage icon beneath Effective Threshold for a particular datapoint:
In the manage thresholds window you can add new thresholds for this DataSource instance that will override the thresholds inherited from the DataSource definition:
Consider the following CPU graph. A threshold was set for the CPU Busy Percent datapoint at the instance level (only this DataSource instance gets the threshold), and to avoid an alert for just a spike occurrence the alert trigger interval was set to 4 at the global DataSource level (all CPU Busy Percent datapoints in the account have an alert trigger interval of 4). Then the CPU exceeded the specified threshold for more than four consecutive cycles and an alert was triggered on the fifth cycle:
The following demonstration shows how to tune alert thresholds at the global and instance levels: