Controlling Throughput on API
The infrastructure that the API services are deployed has a cost factor coined with it and relative to resource consumption on infrastructure such as CPU, Memory, Disk Space, Bandwidth etc. the cost factor of infrastructure would also increase.
A high volume of calls on API services at at given time would increase server resource consumption and if auto-scaling is enabled over a cloud infrastructure, it will spin-up more instances which will increase the cost of infrastructure. If the load is due to an unauthorized intrusion it would be a chaos. Thus, use of a limited rate-for-service calls during a specific time block is important in API communication.
Limiting of API service consumption is generally applicable for concurrent calls made by a single consumer who is identifiable with a single API identity. The thresholds of API throttling has to be set based on the predicted consumption of…
View original post 2,990 more words