Handling 1 million requests per minute


I’ve ran into this blog again and reminded me that this must be one of my favourite blogs on Go : Handling 1 million requests per minute

I’m amazed at what a difference applying a basic pattern can make and the simplicity of the code which would be much more complex implementation in most other languages.

I still have some questions about it though.

What would be good practice numbers to use for MaxWorkers and MaxQueue? Will MaxWorkers map to the number of CPU’s/cores the machine is running?

Also, it all still depends on how quick the call to S3 takes. If the call takes long enough then a backlog in the queue will still be possible, no matter how many workers are running, am I right? Would a timeout to S3 be a good way to prevent this for instance?


This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.