:Scheduled events for database services

Event notifications

Notifications for scheduled events are typically sent 1 to 3 days in advance by email based on your settings in Message Center. To receive these notifications, ensure that the correct recipients are configured in your Alibaba Cloud account and that subscriptions are enabled for ApsaraDB Fault or Maintenance Notifications > Message Settings > Common Settings. After receiving a notification titled "Notification of ApsaraDB for Planned operation and maintenance", you can view details on the Scheduled Events page in Event Center (or Event Management). These details include the event type, region, reason, cancellation risk, and affected instances. You can also adjust the scheduled switchover time of an instance to its off-peak hours as needed.

Notes

• Events are classified into two levels of urgency:

  • [S0 Urgent] Risk Remediation: Typically involves unexpected issues requiring immediate action to prevent service failures. Notifications for these events are sent three days or less in advance, and the window for rescheduling the switchover time is shorter. Examples include urgent version replacement and upgrades, host failure recovery, and upgrades for expiring SSL certificates.

  • [S1 Scheduled] System Maintenance: Involves low-risk issue fixes or planned software and hardware upgrades. Notifications for these events are usually sent more than three days in advance, and you can cancel the event.

• To receive notifications for scheduled O&M events, log on to the Message Center, and select the notification methods and configure contacts for ApsaraDB Fault or Maintenance Notifications. We recommend adding your database O&M personnel as contacts. Otherwise, you will not receive event notifications.Notification methods include Email and Internal Messages. We recommend you select Email to ensure delivery.

  In the upper-right corner of the console, click the bell icon. Then, in the Internal Message Notifications panel, click Message Settings.

  Figure 1. Accessing Message Settings

  

  Figure 2. Notification settings for ApsaraDB

• To receive real-time updates on O&M events or implement custom, event-driven automation, configure system event subscriptions in CloudMonitor. ApsaraDB pushes system events to CloudMonitor for each stage of the O&M event lifecycle, such as Scheduled, In Progress, Succeeded, and Canceled. For more information, see Manage event subscriptions (Recommended). For a list of subscribable CloudMonitor events, see Appendix 1: CloudMonitor-related system events.

  Expand to view a sample CloudMonitor event.

  {
    "eventId": "c864b30b-7f69-5f04-b0e7-8dfb0eabcfd9", // A unique ID for the event.
    "product": "RDS",                                  // The service code.
    "reason": "Host software/hardware upgrade",        // The reason for the event.
    "extra": {
      "impactEn": "Transient instance disconnection",  // The impact of the event.
      "eventCode": "rds_apsaradb_transfer",            // The code for the O&M event type.
      "eventNameEn": "Instance migration",             // The name of the O&M event.
      "switchTime": "2024-09-15T01:30:00+08:00",       // The scheduled switchover time. A transient instance disconnection may occur at this time.
      "startTime": "2024-09-14T21:30:00+08:00",          // The time when the event is queued for execution.
      "cancelCode": "OutOfGoodPerfBySoftHardwareUpgrade", // The cancellation risk code. For details, see Appendix 2: Detailed cause codes and cancellation risks.
      "detailCode": "HostSoftHardwareUpgrade",            // The detailed cause code. For details, see Appendix 2: Detailed cause codes and cancellation risks.
      "instanceInfo": ""
    },
    "instanceId": "rm-2ze9d66o65q1g02g6",             // The instance ID.
    "eventType": "Maintenance",
    "instanceComment": "rm-2ze9d66o65q1g02g6",        // The alias of the instance.
    "instanceType": "Instance",
    "publishTime": "2024-09-10T16:01:47+08:00"
  }

Procedure

1. Log on to the console for your database product.

2. In the navigation pane on the left, click Event Center (formerly known as Event Management) > Scheduled Events. In the top navigation bar, select the region where your instance is located.

3. The Scheduled Events page shows the details of scheduled events. By default, events with the Planned status are displayed. To view historical events, click the Completed or Canceled tab. The following table describes the event attributes.

   Attribute	Example	Description
   Event type	Risk remediation	Events are classified by urgency into two types: Risk remediation and system maintenance.
   Status	Pending	The scheduling status of the event. Note the following key statuses: Waiting Setting Time: The execution time is not set. You must set a time based on your business requirements. If you do not set a time by the Deadline, the event is automatically canceled. Pending: The event is waiting to enter the scheduling phase at the scheduled Start time. Executing: The event is in progress. You cannot manually intervene at this stage. To request an emergency stop, submit a ticket. Note that such non-standard operations can pose unknown risks. Successful: The event has completed. Canceled: The event execution failed or was canceled. Common reasons for cancellation include: Canceled by user (UserCancel): The user canceled the event in the console or by calling an API operation. User response timeout (UserResponseTimeout): The event was automatically canceled because the execution time was not set before the Deadline. Canceled by control system (SupervisorCancel): The event initiator canceled the event. Canceled as unnecessary (AvoidCancel): The event is no longer required because the risk has been resolved or the instance's current state makes the event obsolete. For example, the instance is already on the latest version and requires no update. Canceled by system (AutoCancel): The system periodically inspects scheduled events. The event is canceled if the instance does not meet the execution prerequisites, for example, if the instance is in an abnormal state and cannot receive commands. Execution timeout (ExecuteTimeout): The event entered the execution queue but did not finish within the expected time. Execution failure (ExecuteFail): The event failed during execution due to an unknown error.
   Event type	minor version update	The specific name of the event type. For more information, see Event types and impacts.
   Cause	-	For more information, see Appendix 2: Detailed cause codes and cancellation risks.
   Business impact	Transient connection interruption	The business impact varies depending on the event. For more information, see Event types and impacts.
   O&M suggestions	Ensure your application has an automatic reconnection mechanism and is designed to handle the potential business impact.	O&M suggestions vary depending on the event. For more information, see Appendix 1: CloudMonitor-related system events.
   Start time	-	The time when the event enters the scheduling queue. Before the start time, the event does not affect the instance. After the start time, your database remains accessible, but you cannot perform instance-level operations such as changing the configuration or migrating across availability zones. This field is empty if the event status is Waiting Setting Time.
   Scheduled switchover time	-	The time when a primary/secondary switchover or link switchover occurs, if applicable. This is typically when a transient connection interruption happens. The switchover is expected to occur around this estimated time. In rare cases, such as a failback to the original availability zone, a second switchover may occur. Note A preparation period is required before the switchover for tasks such as event scheduling and data preparation. Therefore, a time gap exists between the start time and the switchover time. This gap varies by database product and event type.
   Deadline	-	The latest time by which you can set the switchover time. The adjusted switchover time cannot be later than this Deadline.
   Cancelable	Yes	You can cancel the event to prevent it from running. This option is typically available for "system maintenance" events. Important Scheduled events are typically issued by the cloud database management system during regular inspections. If you cancel an event, a new one may be issued in the next inspection cycle. Frequent cancellations can increase risks. To avoid these risks, schedule the event for a suitable time based on your business needs instead of canceling it. For more information about cancellation risks, see Appendix 2: Detailed cause codes and cancellation risks.
   Reschedulable	Yes	You can change the execution time for most events. In rare cases involving urgent, high-risk fixes, the time window may not be long enough to allow for rescheduling.

4. (Optional) Modify a scheduled event.

   Select the event that you want to reschedule and click Modify Scheduled Event. You can modify the event in one of two ways:

   • Immediate execution: The event's start time is set to the current time, and it enters the execution queue immediately.

   • Switchover at a specified time: Select a suitable time for the switchover from the available range. The start time is automatically calculated based on the switchover time. The new start time cannot be earlier than the current time. Otherwise, the modification fails.

5. (Optional) Modify the recurring time window.

   In the upper-right corner of the event list, click Recurring Time Window Settings to open the configuration page.

   The execution time of a scheduled event is typically calculated based on the instance's maintenance window. For more information, see the documentation about setting the maintenance window for RDS, Tair, MongoDB, and PolarDB. You can also define a custom recurring time window to meet your O&M requirements. The cloud database service then prioritizes this custom window when scheduling new events.

   You can set the window on a weekly or monthly basis. For example, if you set the recurring switchover time to 02:00–03:00 on every Monday and Tuesday, and the platform's overall event window is from this Tuesday to the following Sunday, the system will identify both 02:00–03:00 on this Tuesday and 02:00–03:00 on next Monday as valid switchover times. Typically, the earlier time (this Tuesday) is prioritized.

   Important

   1. This setting applies only to new events. To change the time of an existing event in the list, click Configure Execution Time.

   2. This setting is only an aid for calculating the execution time and applies only to "system maintenance" events. The actual execution time is subject to the time displayed in the event list.

   3. This is an account-level setting. Once configured, it applies to all database products that support recurring time windows.

6. (Optional) Cancel a scheduled event.

   Select the event you want to cancel and click Cancel Scheduled Event. After reviewing the risks, click Confirm to cancel the event.

Event types and impact

Appendix

Appendix 1: CloudMonitor system events

For more information, see Supported cloud services and their system events.

Appendix 2: Reason codes and cancellation risks

Appendix 3: Event types