System events

Overview

After you purchase your required ECS instances, complete their initial configurations, and deploy your services, maintaining the health of your ECS instances is crucial to the running of your services. To guarantee system performance and the security of your assets, ECS regularly reviews the health of the physical servers where your ECS instances reside. If ECS detects any hardware/software faults or potential risks on the physical servers, your instances will be migrated to healthy servers in real time. During such maintenance tasks, your instances will not be affected and you will not receive any notifications.

If a system event occurs, ECS will send you a notification, information about the recommended handling actions, and the time at which ECS is scheduled to perform these actions. For scheduled system events, ECS will send a notification to you that includes information such as the impact of the system event on your instances and the expected execution time. To prevent any detrimental impacts to your services, we recommend that you back up your data and redistribute inbound traffic before handling system events. After a system event is handled, you can query the historical system events of the past seven days for further analysis and review.

Limits

Phased-out instance type families, such as sn1, sn2, t1, s1, s2, s3, m1, m2, c1, c2, c4, ce4, cm4, n1, n2, and e3, do not support system events. For more information, see Instance type families.

System event types

The following table describes the types of ECS system events.

Event statuses

The following table describes the statuses of a system event during its lifecycle.

System event periods

System events observe the following two periods:

• User operation period: The period between the time when a system event is initiated and the time when the system event starts to be executed. For maintenance-related system events, the period is between 24 and 48 hours. For Subscription instances that are about to be stopped due to expiration, the period is 3 days. For Pay-As-You-Go instances that are about to be stopped due to overdue payment, the period is less than 1 hour. Instances in which a system event occurs due to billing issues will be immediately stopped, and will be released 15 days later if the issue is not resolved.

  During this period, you can use the recommended methods to handle system events in advance, or wait until the default actions are executed. If ECS fixes a system failure and triggers a system event, ECS will send you an event notification in advance based on the system maintenance schedule.

• System action period: If you do not handle a system event in advance, the system event will be automatically fixed within 6 hours after the system action period begins at a scheduled time. After that, you will receive a system event report.
  
  
  Note Only scheduled system events have a user operation period. Unexpected system events that are caused by emergency failures or invalid operations do not have a user operation period. If an unexpected system event occurs, you will receive a notification, but you cannot take any actions. You can only query the historical system events for fault diagnosis and data recovery.

View system events

Procedure (through the console)

If a system event is scheduled, the Unsettled Events button in the ECS console will display a notification badge with the current number of unsettled events.

1. Log on to the ECS console.
2. In the left-side navigation pane, select Overview.
3. In the Common Operation area, click Unsettled Events.
4. Click the Scheduled system event button. A list of instance IDs, regions, statuses, event types, and recommended actions is displayed on the page. Alternatively, you can choose recommended actions from the Actions column to handle system events.

In this example, Alibaba Cloud CLI is used as the tool to call the API. For more information, see Quick start for ECS APIs.

1. Obtain the instance ID.

   aliyun ecs DescribeInstances --RegionId <TheRegionId> --output cols=InstanceId,InstanceName

2. Call DescribeInstancesFullStatus to view the system events of the instance.

   aliyun ecs DescribeInstancesFullStatus --RegionId <TheRegionId> --InstanceId.1 <YourInstanceId> --output cols=EventId,EventTypeName

Procedure (through instance metadata)

For more information, see Metadata.

Modify the scheduled restart time

In the ECS console, you can modify the execution time of system events or manually restart the instance before the user operation period. These operations only apply to system events that have a scheduled restart, for example, an instance is restarted due to system maintenance. To modify the scheduled restart time, follow these steps:

1. Log on to the ECS console.
2. In the left-side navigation pane, click Overview.
3. In the Common Operation area on the Overview page, click Pending Tasks.
4. On the System Tasks tab page, select the target region.
5. Locate the instance that has a scheduled restart, and then click Schedule Restart in the Actions column.
6. In the displayed Scheduled Restart Time dialog box, set the Scheduled Date and the Scheduled Time.
   Note The scheduled restarted time cannot be later than the Latest Schedule Time, which is 30 minutes earlier than the time displayed in the Executed At column.
   
   
   
7. Click OK.

View historical system events

You can query the historical system events of the past seven days for fault diagnosis and review.

Procedure (through the console)

1. Log on to the ECS console.
2. On the left-side navigation pane, click Overview.
3. In the Common Operation area, click Unsettled Events.
4. In the left-side navigation pane, click All events, and then choose Scheduled system events > Instances. A list of instance IDs, event types, and task statuses is displayed.

Procedure (through an API)

1. Obtain the instance ID.

   aliyun ecs DescribeInstances --RegionId <TheRegionId> --output cols=InstanceId,InstanceName

2. Call DescribeInstanceHistoryEvents to view the system events of the instance.

   aliyun ecs DescribeInstanceHistoryEvents --RegionId <TheRegionId> --InstanceId.1 <YourInstanceId> --output cols=EventId,EventTypeName

Subscribe to event notifications

You can set alarm rules for all system events by using CloudMonitor so that you can receive notifications when a system event occurs. For more information, see Cloud product system event monitoring.

Handling suggestions

You can handle system events by using the ECS console or using the corresponding APIs. You can also optimize instance performance based on system events. The following table describes the recommended actions to help you handle system events as needed.