Rate limiting
Capping how many requests a service will accept per tenant or per caller, so one consumer cannot exhaust a shared deployment, whether enforced by a policy in front of the application or inside its own process.
Overview
Present, not yet described.
Inside the CAP process, where tenant and user are already resolved, cds-rate-limitcds-rate-limitA CAP plugin that applies a rate-limit pattern inside the CAP Node.js runtime, capping how many requests a caller gets where tenant and user are already resolved rather than at an approuter or gateway.independent, cap and the rate-limiting feature of @cap-js-community/common@cap-js-community/commonSix independent features for CAP services in one package: replication cache, migration check, rate limiting, a Redis client broker, a local HTML5 repository and a CDM builder.community org, cap cap requests per service and tenant. In front of the application, API ManagementAPI ManagementAPI proxies in front of backend services, with security and traffic management policies attached to them, and Developer Hub where consumers subscribe.sap official, btp from SAP applies its quota and spike-arrest policies on a proxy.
On BTP rate limiting is a policy on an API proxy in front of the application, configured in API ManagementAPI ManagementAPI proxies in front of backend services, with security and traffic management policies attached to them, and Developer Hub where consumers subscribe.sap official, btp and enforced before a request reaches the application's process.
API ManagementAPI ManagementAPI proxies in front of backend services, with security and traffic management policies attached to them, and Developer Hub where consumers subscribe.sap official, btp limits requests with two policies on an API proxy: Quota, a counter per period, and Spike Arrest, a minimum spacing between requests. The apim-as-route-service plan binds the proxy to a Cloud Foundry application's route, so the limit applies whichever URL is called.
Not covered: It stops at the proxy: a call that does not pass through API Management is not counted, and the route-service binding exists for Cloud Foundry applications only. Spike Arrest is kept per API Management server and is not distributed, so its limit is approximate.
No community artifact is documented here.
Artifacts
Content
Show 3 more
Details
Coverage
For your agent
- Ref
capability:rate-limiting- JSON
- /capabilities/rate-limiting.json
- MCP
- Connecting an agent