9.3.1. Creating Monitoring Alarm¶

This section describes how to create Monitoring Alarms.

To create a new Monitoring Alarm, select [ Alarm ] from Monitoring Control Panel and Click [ Create ] or [ Bulk Create ] on the Alarms section screen.

At the Create Alarm screen of Monitoring, the following 4 items are required to be defined:

Basic Information of your alarm	Enter your alarm name (Alarm Name), and its details (Description and Severity).
Conditions of your alarm (Threshold)	Define the relevant thresholds for each resource and meter.
Action for the alarm occurrence	Define conditions for of the action (exceeding or falling below the relevant threshold), the type of action (Email notification or API posting), and details of the action (Email address receiving the notification or API request sentences).
Advanced settings	Define necessary rules for making statistics on the alarm. At the Advanced Settings, the following 5 items are available to be defined: Period ･･･ specify the scope of statistic information to be generated by seconds. Statistics ･･･ Select a statistics type adapted （count/sum/avg/min/max）. Evaluation Periods ･･･ Specify the number of evaluations made on the acquired statistics data. Evaluation Type ･･･ Specify the action execution conditions. (Select “and” for executing the specified action when all of each evaluation meets conditions. Select “or” for executing the specified action when more than two evaluations meets conditions.) *Repeat Actions ･･･ Choose either on or off for making the alert repeatedly. If you want to execute the alert when a sample value exceeds or falls smaller than the relevant threshold more than once, this Advance Settings is not required.

Note

At the Advanced Settings, the scope of samples for the statictics is from a sample value that exceeds the relevant threshold at the timing of settings made to previous sample values made in the relevant number of evaluations performed in the past.

Therefore, even when you set the period of statistics shorter than 5 minutes and specify the evaluation period on the meter such as the usage rate of CPU, whose sample record interval is set every 5 minute, it takes at least the time equivalent to 5 minutes x the number of evaluations (Evaluation Periods) until the stataus of the meter becomes on alarm.

Note

Keep-alive monitoring shows 0 for normal or 1 for outage.

Note

The number of actions taken for alarming is 5 at maximum.

The following sections show examples of settings for creating alert to notify Virtual server instance (CentOS_Tutorial)’s CPU exceeds the threshold or to notify its instance is stopped.

9.3.1.1. [Examples] Creating Alarms for High CPU Usage¶

The following shows an example of alarm settings which “send an email notification when the CPU usage excesses the threshold defined at 80%.”

Define a condition that “an alert will be made when the CPU usage becomes larger than 80%” and an action that “to send email notification when the alert is on” which will be taken in case the defined condition is met.

Note

In the example above, if you want to receive the notification email when the alarming status turns from ALARM to OK (when the CPU usage becomes lower than 80%) as well, add an action taken for the status becomes OK at the section “3. Define Your Alarm Actions”.

Note

At the section “3. Define Your Alarm Actions”, please make sure to click [ +Add ] after inputting all necessary information so that the action made is added to the Added Actions list.

In addition to the example above, another example of alarm settings that “make email notification when the average CPU usage acquired within the past 10 minutes exceeds the threshold defined at 80% at 6 evaluations in a row”.

This defines a condition that “an alert will be made when all of each 6 statistics value, which are used for culculating the average of the CPU usage over 600 seconds, exceeds 80 % of the threshold”.

Note

In this example, the scope of samples for the statictics is from a sample value that exceeds the relevant threshold at the timing of settings made and previous 6 sample values made in the past evaluations. As each sample value is recorded every 10 minute, it takes at least 60 minutes to be the meter on the ALARM status.

9.3.1.2. [ Examples ] Creating Alarms for Instance Shutoff¶

The following shows an example of alarm settings which “send an email notification when the target virtual server instance is shutoff.”

Define a condition that “the ALART status means when the meter of the Virtual Machine Status is not zero (0, unbooting status)” and an action that “to send email notification when the ALARM status is on” which will be taken in case the defined condition is met.

Note

At the meter of the Virtual Machine Status, zero (0) means the instance is booted.

Note

In the example above, if you want to receive the notification email when the alarming status turns from ALARM to OK (when the instance is back to booting status) as well, add an action taken for the status becomes OK at the section “3. Define Your Alarm Actions”.

Note

At the section “3. Define Your Alarm Actions”, please make sure to click [ +Add ] after inputting all necessary information so that the action made is added to the Added Actions list.

9.3.1.3. [ Example ] Bulk Creating Alarms for High CPU Usage and Instance Shutoff¶

With the Bulk Creating Alarms function, you can create multiple alarms to a single resource and/or a single alarm to multiple resources.

For the example here, create the alarms to multiple virtual server instance (Ubuntu_Tutorial and RHEL_Tutorial) that “send an email notification when the CPU usage becomes larger than 80%” and “send an email notification when the target virtual server instance is shutoff” together.

9.3. Using Monitoring Alarms

9.3.2. Managing Monitoring Alarms (Editing and Deleting)