Dear Tencent Cloud user,
The Service Level Agreement (SLA) for TencentDB for SQL Server was updated on August 14, 2026. Please see the details below.
Effective Date
August 14, 2026 (Friday), Beijing Time.
Update Details
In Definitions, Section 1.6: "Monthly Service Fee" has been updated:
Previous | Updated |
The total amount of service fees that you pay for a single TencentDB for SQL Server instance in a service month, excluding the portion that has been purchased but not consumed. If you pay for more than one month at once, the monthly service fee will be amortized based on the number of months purchased. | Monthly Service Fee refers to the total service fees you actually pay for a single instance within a service month, excluding vouchers, credits, discount deductions, complimentary allowances, and any prepaid but unconsumed portions. If you pay for multiple months of service fees in one lump sum, the Monthly Service Fee is calculated by prorating the total payment based on the number of months purchased. |
In Exceptions to this SLA, Section 4.2 has been added:
4.2 Once an admin account (an account with SA permissions) is created for an instance, including scenarios where it is created and subsequently deleted, sysadmin permissions are enabled. Instances for which an admin account has been created will no longer enjoy SLA protection. Database instances that have never created an admin account are not affected. After an admin account is created for an instance, admin account holds excessive permissions and the operational permissions it holds will exceed the control scope of TencentDB for SQL Server management, therefore the service availability, control availability, data availability, data reliability, and data security of the instance cannot be guaranteed.
In Limitation of Liability, Section 5.11 has been updated:
Previous | Updated |
temporary interruption or suspension of the Service due to the "routine maintenance" described in the service documentation for TencentDB for SQL Server. | Service unavailability time does not include brief service interruptions or suspensions caused by Tencent Cloud performing "routine maintenance" (including but not limited to system maintenance, version upgrades, hardware replacement, data center migration, disaster recovery switchover drills, and simulated failure drills, and so on) as stipulated in the Tencent Cloud Terms of Service for the TencentDB for SQL Server service, provided that advance notice is given or the maintenance is conducted within the agreed maintenance window. |
In Limitation of Liability, Sections 5.14, 5.15, 5.16, and 5.17 have been added:
5.14 You shall promptly take action to perform emergency service upgrade or version upgrade upon a notification issued by this service's technical services. If you fail to perform the upgrade within six months of being notified due to your own inaction, you are solely responsible for any resulting losses and consequences, such as business interruption or data loss.
5.15 To ensure data security, you are responsible for backing up your data. Tencent Cloud products or services may provide you with tools that feature daily data backup. However, backup or restoration may fail due to factors such as lock conflicts between certain DDL operations and the backup process during backup, insufficient disk space or a full volume, long-running large transactions preventing transaction log truncation, improper configuration of the recovery model (for example, changing from Full to Simple) preventing log backup, and the disabling or deletion of certificates/keys required for encryption affecting subsequent restoration. As these factors involve customer business operations, which Tencent Cloud cannot address without authorization, Tencent Cloud cannot guarantee successful backup under all circumstances.
5.16 If backup or rollback fails, or backup data is missing or unavailable due to the following user-side reasons, such cases are not considered service unavailability, and Tencent Cloud is not liable for compensation, including but not limited to:
(1) Insufficient disk space or a read-only/locked state of the instance disk prevents full or incremental backups from being written.
(2) The presence of long-running active transactions or uncommitted transactions (ACTIVE_TRANSACTION) prevents transaction log truncation, resulting in continuous log growth, log backup failure, or the inability to complete backups within the agreed timeframe.
(3) After SA is enabled for an instance, user operations such as manually truncating logs, skipping or deleting intermediate log backups, or changing the recovery model can break the log chain, preventing continuous point-in-time restoration.
(4) After SA is enabled for an instance, performing file management operations (such as ALTER DATABASE to add/remove data or log files) or shrinking databases/files (including autoshrink) within the backup window can cause lock conflicts with the backup process. If these conflicts result in lock wait timeouts, the backup will fail.
(5) If the certificates or asymmetric keys on which an encrypted database (such as TDE) depends are disabled, deleted, or become inaccessible by the user, subsequent restoration of the backup cannot be performed.
(6) Other backup or restoration exceptions caused by user-controllable factors such as business operations, resource planning, table design, or key management.
5.17 Service unavailability time does not include performance degradation or brief unavailability caused by the instance's own load reaching resource bottlenecks, such as slow response, hang, connection exhaustion, resource overutilization (CPU, memory, IOPS, disk, bandwidth, and so on), primary-replica replication delay or interruption, or performance issues during large transactions, DDL operations, or log replay.