> If the threshold is too large, we still can't find it if there is a problem. 
> So how do we determine a suitable threshold?
> I think this is difficult. And even if we find this threshold, ut can't be 
> stable, so I don't think we should handle this.
> 
> How do u think about it?

Currently the threshold is 12% diff to expected requests should dispatch. Maybe 
there's some problem in Random for example, it will between 0~8% (I think the 
result should be stable when we simulate 10k request).

12% under weight is acceptable to me(more like a complete unit test and 
reliable after hundreds loop), but your opinion is reasonable too.

So i won't consist on it and we can make a decision whether to leave them as 
they were.

btw the CI failed and i am checking the problem(seems not because this unit 
test).

[ Full content available at: 
https://github.com/apache/incubator-dubbo/pull/2647 ]
This message was relayed via gitbox.apache.org for 
[email protected]

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to