Effected time May 7, 2025 00:00:00 - May 7, 2025 06:00:00 (UTC+08) Changes To improve service quality and system stability, the infrastructure and network team will perform network equipment upgrades on May 7, 2025 (UTC+08). The changes are Software maintenance/upgrade, cold patch. Effected scope Hong Kong Zone D Description The upgrades will cause 2 brief network jitters/delays/packet losses, each lasting around 3 seconds. Do not perform business operations during this time. We apologize for any inconvenience.

Change Content Edge Node Service(ENS) snapshot and standalone cloud disk currently bill as pay-by-monthly-peak. To better meet customer needs, ENS snapshots and standalone cloud disks will now support pay-as-instance-capacity, you can choose monthly、daily or hourly billing cycles This policy adjustment will take effect on June 1, 2025. Pricing details will be announced before the adjustment. For updates, refer to the storage billing page: https://www.alibabacloud.com/help/en/ens/product-overview/cloud-disk-billing Pay-by-monthly-peak：You are charged based on the monthly peak capacity of the cloud disk. ENS calculates the capacity of the cloud disk every 5 minutes, and uses the maximum value among the 288 results in a day as the daily peak value of memory usage. At the end of a month, the maximum value among the daily peak values is used for billing. Pay-as-instacne-capacity: Each cloud disk is billed individually, One Cloud disk capacity (GB) × Unit price × Usage duration Impact of Changes The adjustments will affect you as follows: New instances: Newly created snapshot instances and standalone cloud disk instances will be billed based on the billing method you selected. Existing instances: Existing snapshot instances and standalone cloud disk instances will be automatically switched to the Pay-as-instacne-capacity billing method. The adjusted total cost will not exceed the previous Pay-by-monthly-peak billing cost. Example: If you created two Ultra cloud disks on April 1, 2025,Disk-1(50 GB) and Disk-2(100 GB). Then you released Disk-1 on April 15, 2025, while Disk-2 remained in use: Pay-by-monthly-peak billing: Total cost for April 2025 = (50 GB + 100 GB) * 0.0493 USD/GB/month = 7.395 USD Pay-as-instacne-capacity(monthly) billing: Total cost for April 2025 = 50 GB * 0.0493 USD/GB/month * 15/30 + 100 GB * 0.0493 USD/GB/month = 6.1625 USD

During the upgrade period, except for the following scenarios, the release process will be error-free, imperceptible, and request handling will not be interrupted: 1. When the client is deployed on ECS instances within the Classic Network, there might be brief connection interruptions or connection refusals lasting for a few seconds. 2. When using domain name mappings to fixed IP addresses or custom DNS resolutions (e.g., local DNS resolution), there might be brief connection interruptions or connection refusals lasting for a few seconds. 3. During the release period, if there is a message backlog in the queue, requests for deleting messages or modifying message visibility time may result in client errors. For the above scenarios, please ensure that your client implements reconnection and retry mechanisms. We apologize for any inconvenience this may cause. If you have any issues, please feel free to contact us via a support ticket.

Charging Time: 09:00 on June 10, 2025, Beijing Time Charging Scope: Windows operating systems used on nodes in Hong Kong, Macau, Taiwan of China and overseas nodes Adjustment Content: Alibaba Cloud will officially charge for the Windows operating system, a mirror image fee will be charged for using this type of operating system. Charging Rules: Existing and newly created instances using the Windows operating system on nodes in Hong Kong, Macau, Taiwan of China and overseas nodes; On-demand pay-as-you-go instances of the activation type will no longer support the Windows operating system. For specific price details, please refer to: https://www.alibabacloud.com/help/en/ens/product-overview/mirror-billing

Upgrade a router device in Malaysia , during which traffic migration and restoration operations are involved. Upgrade impact: The upgrade could impact the internet access of some users in Malaysia During the upgrade, route convergence may cause two minute-level jitters. Jitter has no material impact on long-distance connection services with automatic reconnection function. Please make sure that you have a reconnection and retry mechanism placed in your business to enhance the fault tolerance of your business. We apologize for the inconvenience caused to you. If you have any questions, please get in touch with us for consultation and feedback.

To enhance the stability of data queries, starting from 2025-05-26 00:00:00 (UTC+8), the Observable Monitoring Prometheus Edition will impose a maximum time span limit of 30 days for non-aggregated queries under Remote Read and GlobalView aggregated instances across all instances. Note: Aggregated queries refer to metric query requests that use operators such as sum / avg / min / max and do not include GlobalView-specific labels (unique_cluster_id (instance unique identifier), unique_cluster_name (instance name), unique_cluster_account (account type), unique_cluster_region (instance region)) in the conditions. We apologize for any inconvenience this may cause. If you have any questions, please feel free to contact us through a support ticket or our service hotline.

Upgrade content: upgrade of network equipment located in Alibaba Cloud Riyadh node. Upgrade impact: Users accessing Alibaba Cloud node in Riyadh through SaudiNet might be impacted. Under normal circumstances, the users have no perception on the impact. In abnormal cases, route convergence may cause two minute-level jitters. The jitter basically has no effect on the long-connection service with automatic reconnection function. Please ensure that you implement the reconnection and retry mechanism in your services to enhance the fault tolerance of your services. We apologize for the inconvenience caused to you. If you have any questions, please contact us for consultation and feedback.

Dear Alibaba Cloud Model Studio User, We will update the Model Studio Section of the "Alibaba Cloud Product Terms." You can find the updated terms in Article 4.48 via the following link: https://www.alibabacloud.com/help/en/legal/latest/alibaba-cloud-international-website-product-terms-of-service-v-3-8-0. The updated terms will take effect on May 8, 2025. If you have any questions, please feel free to contact us. Thank you for choosing Alibaba Cloud Model Studio!

Upgrade content: upgrade of network equipment located in Alibaba Cloud Querétaro node. Upgrade impact: The upgrade could impact the internet access of some users in Querétaro During the upgrade, route convergence may cause four minute-level jitters. Jitter has no material impact on long-distance connection services with automatic reconnection function. Please make sure that you have a reconnection and retry mechanism placed in your business to enhance the fault tolerance of your business. We apologize for the inconvenience caused to you. If you have any questions, please contact us for consultation and feedback.

Update content ApsaraMQ for RocketMQ in the corresponding region, including access methods over TCP, MQTT, and HTTP Upgrade Impact During the update, the following issues may occur: 1. Second-level transient connections may occur in the ApsaraMQ for RocketMQ console and each service node of the cluster. The transient connection duration is positively correlated with the scale of the cluster. In most cases, transient connections do not affect business because the automatic retry mechanism is triggered on clients. However, exception logs are generated and message delays may occur. 2. Messages may be duplicated. Make sure that message idempotence is performed on applications based on best practices. 3. During the update, do not manage resources such as creating or deleting topics in the ApsaraMQ for RocketMQ console. 4. If you access ApsaraMQ for RocketMQ over HTTP, transient connections may occur or connection requests may be rejected. The issue lasts for less than 1 minute. Make sure that the reconnection and retry mechanisms are configured on clients. You can also use the monitoring feature provided by ApsaraMQ for RocketMQ to monitor important business. To configure the monitoring feature, go to the ApsaraMQ for RocketMQ console and click Monitoring and Alerting in the left-side navigation pane of the Instance Details page. We apologize for the inconvenience. If you have questions, submit a ticket to contact ApsaraMQ for RocketMQ technical support.