Request cost index for throttling requests to execute operations in a multi-tenant provider network

A method for throttling graph query language requests in a multi-tenant provider network, the method comprising:

• receiving, by an application programming interface (API) service in the multi-tenant provider network, a graph query language request to execute an operation, the operation of the graph query language request comprising a set of one or more fields to be resolved to data, the set of one or more fields having a field cardinality;
• determining, by the API service, to process the graph query language request based on a request rate limiting algorithm;
• executing, by the API service, the operation to yield an operation result, wherein executing the operation comprises the API service executing a set of one or more resolver functions associated with the operation to resolve the set of one or more fields of the operation of the graph query language request to data, wherein the executing the set of one or more resolver functions encompasses a number of resolver function invocations, and wherein the number of resolver function invocations is greater than the field cardinality of the set of one or more fields of the operation of the graph query language request;
• determining, by the API service, a respective request computing resource cost index for each resolver function of the set of one or more resolver functions, the respective request computing resource cost index reflecting an amount of computing resources utilized by the API service executing the resolver function;
• calculating, by the API service, an overall request computing resource cost index reflecting an amount of computing resources utilized by the API service executing the operation, the overall request computing resource cost index calculated based on the respective request computing resource cost index determined for each resolver function of the set of one or more resolver functions;
• determining, by the API service, an adjustment amount for a state variable of the request rate limiting algorithm based on the overall request computing resource cost index;
• adjusting, by the API service, the state variable of the request rate limiting algorithm by the adjustment amount;
• and
• sending, by the API service, the operation result.

A system for throttling graph query language requests in a multi-tenant provider network, the system comprising:

• a first one or more electronic devices to implement a gateway service in the multi-tenant provider network, the gateway service comprising instructions which when executed cause the gateway service to:
• receive a graph query language request to execute an operation comprising a set of one or more fields having a field cardinality,
• determine to process the graph query language request based on a request rate limiting algorithm,
• send the operation to an execution service in the multi-tenant provider network,
• adjust a state variable of the request rate limiting algorithm based on an overall request computing resource cost index that reflects an amount of computing resources utilized by the execution service to execute the operation, and
• send an operation result of executing the operation;
• and
• a second one or more electronic devices to implement the execution service in the multi-tenant provider network, the execution service comprising instructions which when executed cause the execution service to:
• execute a set of one or more resolver functions associated with the operation to resolve the set of one or more fields of the operation of the graph query language request to data, wherein a number of resolver function invocations is greater than the field cardinality of the set of one or more fields of the operation of the graph query language request,
• determine a respective request computing resource cost index for each resolver function of the set of one or more resolver functions reflecting an amount of computing resources utilized by the execution service executing the resolver function;
• calculate the overall request computing resource cost index based on the respective request cost index for each resolver function of the set of one or more resolver functions, and
• provide the request computing resource cost index to the gateway service.