Elastic Compute Service:Use OOS to manage ECS

Scenario

Description

Run event-driven O&M tasks

You can use OOS to trigger an O&M action when an event occurs. This way, O&M tasks can be automatically and efficiently run without the need for human intervention.

For example, you can use OOS to trigger instance restart when a high vCPU utilization event occurs. When the vCPU utilization of an ECS instance reaches 85%, the instance is automatically restarted.

Run batch O&M tasks

You can use OOS to run O&M commands on multiple objects, such as ECS instances, at a time to perform regular operations and maintain the normal operation and health of your business.

For example, to check the available space of disks on multiple ECS instances, perform the following steps: Search for ECS instances by instance name, tag, or resource group, run a Cloud Assistant command to check the available space of disks on the instances, and then view the check result.

Run O&M tasks subject to approval

In many scenarios, you must add an approval process to ensure that only secure and expected O&M operations are performed. You can define an ACS::Approve action in a template to ensure that O&M tasks can be run only with manual approval and that only expected O&M tasks are run without mistakes.

Run scheduled O&M tasks

You can use OOS to run O&M tasks based on a schedule. For example, you can use OOS to start, stop, and restart ECS instances on a periodic basis to reduce costs.

Run cross-region O&M tasks

High-availability architectures allow multi-region deployment, which results in increasingly complex O&M tasks. You can use OOS to create templates for O&M operations in different regions to run O&M tasks across regions.

Perform O&M in multiple regions

If you use resources in multiple regions, synchronization is required to maintain consistency among the resources. For example, you may need to enable the logging feature for Object Storage Service (OSS) buckets in multiple regions at a time.