Real time multi-tenant workload tracking and auto throttling

A computer-readable storage medium having computer-executable instructions stored thereupon which, when executed by a processor, cause the processor to:

• receive a plurality of requests from a plurality of tenant computing devices associated with a plurality of tenants;
• process the plurality of requests in part by sending a second plurality of requests to an underlying resource;
• determine a resource usage of the underlying resource attributed to each of the plurality of tenants by determining one or more usage metrics of the underlying resource caused by each of the second plurality of requests and aggregating the one or more usage metrics associated with each of the plurality of tenants, wherein the one or more usage metrics are determined from responses to the second plurality of requests;
• identify an overactive tenant of the plurality of tenants by determining that the resource usage of the underlying resource attributed to the overactive tenant meets one or more overactive tenant criteria;
• receive a request from a tenant computing device associated with the overactive tenant;
• and
• throttle the request, wherein throttling the request comprises including a throttle message in a response to the received request.

A computing device, comprising:

• one or more processors;
• and
• at least one computer storage media having computer-executable instructions stored thereupon which, when executed by the one or more processors, will cause the computing device to:
• receive a plurality of requests from a plurality of tenant computing devices associated with a plurality of tenants;
• process the plurality of requests in part by sending a second plurality of requests to an underlying resource;
• determine a resource usage of the underlying resource attributed to each of the plurality of tenants by determining one or more usage metrics of the underlying resource caused by each of the second plurality of requests and aggregating the one or more usage metrics associated with each of the plurality of tenants;
• identify an overactive tenant of the plurality of tenants by determining that the resource usage of the underlying resource attributed to the overactive tenant meets one or more overactive tenant criteria;
• determine, based on which of the one or more overactive tenant criteria the overactive tenant met, whether to soft-throttle or hard-throttle requests associated with the overactive tenant;
• store in a throttled tenant table a throttle state indicating whether the overactive tenant is soft-throttled or hard-throttled;
• receive a request from a tenant computing device associated with the overactive tenant;
• apply a slow throttle or a hard throttle based on the throttle state.