Effective date: January 15, 2019 This Service Level Agreement (hereinafter referred to as "SLA") sets forth the service availability level index and indemnity plan on the RDS (Relational Database Service, "RDS" for short) provided by Baidu AI Cloud. The user is kindly reminded that this SLA applies only to the RDS's Dual-Compute High-Availability Version, Raft Financial Version, and proxy instances, rather than its Standalone Basic and read-only instances.

1. Definitions

Service cycle: Means a calendar month. Service cycle in minutes: Means the total number of days in one service cycle ? 24(hours) ? 60(minutes). Service Downtime in minutes: If the user's connection attempts for connecting to the designated the RDS instance all fail and such status lasts for more than five minutes, it will be deemed as "Service Downtime" on this the RDS instance within current minute. Accordingly, the total of "downtime" minutes of the RDS instance within a service cycle is called "Service Downtime in Minutes". Monthly service fee: Refers to the total service charges that are paid by the user for one the RDS instance in a calendar month. In case the user has paid service fees of several months at one time, the monthly service fee equals to the lump sum payment divided by the number of months of the Service covered by such payment.

2. Service availability

2.1 Service availability calculation formula

The service availability is calculated as follows with a single instance as dimension:

2.2 Service availability commitments

The Baidu AI Cloud RDS promises that service availability is not lower than 99.95%, meaning that the user's "Service Availability" time per month is 30?24?60?99.95%=43178.4(minutes), namely, there exist downtime minutes: 43200-43178.4=21.6(minutes). If the RDS fails to fulfill the said service availability commitments, the user may claim for indemnity according to Article 3 of this agreement. However, the scope of indemnity does not cover the "Service Downtime" resulting from: (1) System maintenance after Baidu AI Cloud notified users in advance, including handover, maintenance, upgrade and simulated failure drill; (2) Any network or equipment failure or configuration adjustment other than the equipment to which Baidu AI Cloud belongs; (3) Hacker attacks to the user's applications; (4) Time consumed during the RDS instance's backup recovery; (5) Improper setting of the RDS whitelist by the user; (6) The RDS node OOM caused by the complicated slow SQL from the user to the RDS ; (7) Change and deletion to the RDS instance's VPC by the user without approval; (8) Loss or disclosure of data and passwords caused by improper maintenance or confidentiality of users; (9) Negligence of the user or operation authorized by the user; (10) Users' failure to follow Baidu AI Cloud product documentations or use recommendations; (11) Force majeure or other accidents; (12) Unavailability caused by reasons unattributable to Baidu AI Cloud.

3. Indemnity schedule

3.1 Compensation standards

With respect to each RDS instance, the indemnity sum is calculated according to a single instance's monthly service availability and based on the standard in the following table. However, the indemnity mode is only limited to use for the purchase of the RDS product voucher. Moreover, the total compensation shall not exceed the monthly service fees (exclusive of voucher fees deducted) that are paid by the user for the RDS instance in the current month when the service availability commitments fail to be fulfilled.

3.2 Time limit for claim

The user can lodge a claim for the instance that fails to meet the availability standard in the previous month after the fifth (5th) workdays of the current month. However, the claim must be made within two (2) months as from the month in which the RDS instance fails to meet the availability standard. If the user fails to lodge the claim beyond the time limit, the claim will not be accepted.

4. Miscellaneous

(1) Within the scope stipulated by laws and regulations, Baidu reserves the right of final interpretation to this agreement. (2) Baidu AI Cloud has the right to modify the SLA clauses. If there is any change to the SLA clauses, Baidu AI Cloud will notify you by website publicity or email. If you disagree with Baidu AI Cloud's modification to SLA, you have the right to stop the use of the RDS service. If you continue using the RDS service, it is deemed as your acceptance of the modified SLA. (3) All notices given by Baidu AI Cloud SCS Service to users under this agreement can be made by webpage notice, station letter, email, SMS, or in other forms. Such notices shall be deemed to have been delivered to the users on the date of delivery. Baidu AI Cloud shall not be liable to any losses encountered by the users. (4) Formation of this agreement and its execution, interpretation, and settlement of disputes in connection herewith shall be governed by the laws of the People's Republic of China (P. R. C.) and subject to the jurisdiction of the courts of the P. R. C. In case of any dispute between both parties regarding the content of this agreement or its execution, the parties shall try their best to resolve it through friendly negotiation. In case of failure to solve the negotiation by the negotiation, either party may bring a lawsuit to the Haidian District People's Court, Beijing. (5) This agreement constitutes a complete agreement between both parties on the matters agreed in this agreement and other related matters. Except as provided in this agreement, no other rights are granted to any party hereto. (6) If any provisions in this agreement are entirely/partially ineffective or unenforceable for any reason, other clauses in this agreement are still effective and binding upon both parties. (7) For user restriction provisions, see "Users' Rights and Obligations" in the User Service Agreement for more information. (8) For the exemption clause of service providers, please see the relevant clause of "Disclaimer" in the User Service Agreement for more information.