CloudOps Orchestration Service (OOS) lets you create O&M tasks that are triggered by threshold alerts for cloud service metrics. When a metric for a monitored cloud resource reaches a specified threshold, OOS executes a specified template to perform alert-based O&M. These tasks run continuously and monitor for alerts until you cancel them. For example, you can create an O&M task to clear a log directory when disk utilization exceeds 80%.
For more information about supported metrics, see Major metrics of Alibaba Cloud services.
To create an alert O&M task, complete the following steps:
Set an alert rule
Select a template
Set the template parameters for the alert trigger
Set an alert rule
Field | Required | Description |
Product type | Yes | Select a product from the drop-down list. You can select only one. |
Rule description | Yes | The rule for the threshold alert. |
Mute for | No | If the monitoring data continuously exceeds the alert threshold, the alert is triggered only once during each mute period. The default period is one day. |
Effective period | No | The time range during which the alert rule is in effect. By default, the rule is in effect all day. |
Resources to alert on | Yes | Note If you do not select resources using one of the following three methods, the rule applies to all resources in your account.
|
A threshold alert rule includes the following fields:
Metric name
Aggregation period for monitoring data
Number of statistical periods
Statistical method
Comparison operator
Threshold
Select a template
Select the template to execute when an alert occurs.
Set the template parameters for the alert trigger
For template parameters, you can enter static values or select parameters from the alert message. If you use static parameters, the template is always executed with those values. If you select parameters from the alert message, you can configure jq expressions to extract fields from the message.
To extract fields from an alert message, start with a dollar sign ($) and add a jq expression. For example, the following code shows the format of an alert message for the cpu_total metric of an ECS instance:
{
"Average": 50.15,
"Maximum": 50.75,
"Minimum": 49.75,
"curLevel": "INFO",
"instanceId": "i-bp1gn7od******qh5r12",
"ruleName": "alarmtrigger-130920******0047-exec-de81413d******71b537",
"timestamp": 1575970560000,
"userId": "130920******0047"
}To extract the ID of the instance that triggered the alert, use the expression $.instanceId.
The following static fields can be extracted from the alert message:
Extraction expression | Description | Example value |
$.timestamp | The timestamp of the alert, in milliseconds. | 1575970560000 |
$.curLevel | The alert level. | INFO |
$.userId | The ID of the Alibaba Cloud account. | 130920**0047 |
$.dimensionFieldName | The value of the monitoring dimension. Replace `dimensionFieldName` with the name of the metric dimension field. For example, the CPU of an ECS instance is monitored by the `instanceId` dimension. You can use | N/A |
Example of selecting parameters from an alert message:
The method for setting static parameters is the same as for setting regular template parameters.
Notes:
Static parameters (static configuration): These parameters are manually specified when you create the task and do not change during execution. If a task uses these parameters, it always operates on predefined resources and cannot identify the specific instance that triggered the alert.
Alert message parameters (dynamic parameters): These are automatically extracted from the message content of an alert event, such as the
resourceIdandinstanceIdfields in a CloudMonitor event. They accurately reflect the actual resource associated with the current alert.
If a task uses static parameters instead of dynamically obtaining the resource identity from the alert message, the O&M task might operate on the wrong object. This can cause the resource that triggered the alert to not be processed.