Create rate limiting rules to be enforced by on-premises gateways configured to protect your private generative AI services.
Rate limiting rules allow you to control the number of requests endpoints can make
to your protected private generative AI service, helping to prevent denial-of-service
attacks. Rate limiting rules can only be enforced by on-premises gateways operating
in reverse proxy mode and deployed in front of a protected generative AI service.
Procedure
- In Internet Access and AI Service Access Configuration, click Manage Rate Limiting Rules.
- To create a new rate limiting rule, click Create Rule.
- Specify a name for the rule and an optional description.
- Choose whether to enable the rule after creation.
- Select the private IP address groups to target with the rule.
- All private IP address groups: All available private IP address groups are subject to rate limiting when making requests to the private generative AI service
- Any selected private IP address groups: Private IP address groups you select are subject to rate limiting when making requests to the private generative AI service
Tip
You may also choose private IP address groups to specifically exclude from the rule. - Select the on-premises gateways to apply the rule.
- All on-premises gateways: All supported on-premises gateways apply the rule
- Any selected on-premises gateway: Supported on-premises gateways you select apply the rule
Important
Only connected on-premises gateways with reverse proxy mode enabled that are protecting a private generative AI service are supported. The Internet Access On-Premises Gateway service and corresponding Service Gateway must be updated to the latest version. - Specify the rate limit threshold for the rule.
- Unlimited requests: The specified private IP address groups are permitted to make unlimited service requests through the specified on-premises gateways
- No requests: The specified private IP address groups are not permitted to make any service requests through the specified on-premises gateways
- Specified request rate: The specified private IP address groups are permitted to make only the specified number of service requests per minute through the specified on-premises gateways. You may specify a number ranging from 1 to 2048.
- Click Create.The newly created rule appears on the Rate Limiting Rules screen.
- To search for a particular rule, type the rule name in the search field or use the dropdown menu to filter the rules by target or on-premises gateway.
- To enable or disable an existing rule, click the corresponding toggle in the Status column.
- To duplicate a rule, click the copy icon (
). - To delete a rule, click the delete icon (
) or select multiple rules and click Delete.