In Monitis, a multi-value monitor is defined as one that:
- Monitors more than 1 object at the same time
- Monitors the same object from different locations
The idea of multi-value monitor is to have more than 1 check for the same check time.
To configure alerting for a multi-value monitor follow the instructions below, with the example of CPU monitor brought to take you step-by-step through the setup process.
Example: Setting Up Alerting for CPU Monitor
CPU monitor is a multi-value monitor, with cores being the monitored objects. Shown below is the chart view of CPU monitor for Windows, where you can see User and Kernel metrics plotted for the selected core or average across cores.
You can plot on the chart any one of the metrics (User or Kernel) for the selected cores. Go to the Other tab in the monitor settings and select CPU for selected metric.
Back in the chart view you will see that the chart legend has changed: select from the drop-down list the metric to plot CPU values for on the chart.
Setting Up a Threshold in Monitor
In Monitis new Alerts 2.0 alerting model, to set up an alert rule and start receiving alerts whenever your monitor enters a problem state you need first to configure thresholds in your monitor.
Thresholds denote the monitor’s health status. There are two severity level thresholds in Alerts 2.0:
You define conditions for the monitor of entering the Critical or Warning state when adding a new threshold to it. You can define more than one Critical or Warning threshold in a monitor.
See Thresholds for more info.
To add a new threshold in your monitor, go to the monitor settings.
In the Settings window go to the Thresholds tab and click Add Threshold.
The New Threshold window will open.
Select the severity level for the new threshold (CRITICAL or WARNING).
To configure the threshold, first you need to select object(s) to apply the threshold to.
Select any of the CPUs (cores) from the drop down list. You can also select the average across all cores (not all the selected cores, but all cores of the monitored server).
Choose whether the metrics condition in the threshold should apply to “Any”, “All” or “At least X” of your selected objects.
Select from the combo box the threshold condition. Depending on the monitor type, you will see up to 5 options there:
- Result is error
- Result is error or ANY metric matches the condition below
- Result is error or ALL metrics match the condition below
- ANY metric matches the condition below
- ALL metrics match the condition below
For CPU monitor there is no availability check (“Result is Error”), your machine’s CPU is assumed to be always available for measurement provided that Monitis Agent is installed on it (see Downloading and Installing Monitis Agent for Windows and Linux), therefore you will see only the last 2 options in the list.
Now you need to define the metric condition. The metrics can be tied together by either logical AND or logical OR operators – you do that by selecting the “ANY metric” or “ALL metrics” in the combo box above. You cannot use both AND and OR to tie metrics in the threshold.
Select the metric from the list, and set a condition for it by selecting the operator and assigning a value.
Add more metrics to the expression as needed, by clicking the green “+” button.
Click OK when finished.
The newly created threshold will be added in the list of thresholds for this monitor.
Adding Alert Rules
As soon as you have a threshold configured in your monitor, you can set up an alert rule to be notified whenever the monitor enters a problem state.
Go to the Alert Rules tab in the monitor settings, and click Add Alert Rule.
The New Alert Rule window will open.
Select the contacts to send alerts to.
Select the monitor state to alert upon: CRITICAL or WARNING.
If e.g. you select CRITICAL, the alerts will be sent whenever any CRITICAL level threshold is violated in the monitor.
Click Add Rule to finish the setup.
For another detailed example of setting up alerting for a multi-value monitor see SQL Query Result Monitor.