Current limiting and download interface request number control

The cash register system is much more stable than the Internet application traffic , The peak flow is three times as much as usual , It is completely controllable , We use the current limit here to control the background downloading at the same time XLS Number of requests for the table , Because the background system needs to provide a large number of exports XLS Table function , Cashiers and managers need to use the data we export to do various analyses （ It should be because our report function is not perfect ）.

Let's sort it out Guaua Current limiting tools provided and our own practices .

One 、 Use LoadCaching Realize interface current limitation

Ideas ： establish LoadCaching object , Set the validity period to 1 second , cache Key For timestamps （ The current number of seconds ）, Value is a counter （ The number of calls is accumulated ）, If the query hit counter is incremented , Do not process when the counter value is greater than the limit value , If there is no command , Create a new timestamp cache and initialize the value to 0, To achieve the purpose of current limitation .

Two 、 Use RateLimter Current limiting

RateLimter Limit the rate of access to resources , There are two interfaces for obtaining tokens , One is non blocking , If you can't get the token, you can return directly false, The other is blocking , Can't get a token wait for a token .

RateLimiter rateLimiter = RateLimiter.create(10)
 Number of tokens created per second 

3、 ... and 、 The number of download interface requests is limited

Ideas ： Create an interceptor , Use redis Store the currently downloaded request （ Generate a unique string ）, Judge the present redis The key Of size size , If the limit is exceeded , Just return the message you need to wait for , When the download is complete , The unique string of the request is separated from redis Delete in .