I would like to implement a rate limiter based on the time usage of actions.
In a time interval (maybe 15 minutes):
- Every user has
500 ms credits.
- If a user has a
verified trait, then it gets 2000 ms extra credits.
- If a user has other traits, it may get extra credits.
In the above scenario, we can serve around 300 users at once:
$(15 \times 60) \div (0.5 + 2) = 360 \Rightarrow 300$ (We need some resources to do checks and other chores)
Since we do not have extra machines to do scaling and we haven't thought about it, a rate limiter would be necessary.
I do believe that we can scale, but more research must be done.
Feel free to discuss the topic, thanks.
I would like to implement a rate limiter based on the time usage of actions.
In a time interval (maybe 15 minutes):
500ms credits.verifiedtrait, then it gets2000ms extra credits.In the above scenario, we can serve around 300 users at once:
Since we do not have extra machines to do scaling and we haven't thought about it, a rate limiter would be necessary.
Feel free to discuss the topic, thanks.