cserwen commented on issue #3425:
URL: https://github.com/apache/rocketmq/issues/3425#issuecomment-951600335


   > @cserwen @Jason918 Just because Because the timer is single-threaded 
scheduling and is executed sequentially and serially, so another contributor 
has been submitted a PR to polish it, and would you like to help review this 
[PR](https://github.com/apache/rocketmq/pull/3287)?
   
   @duhenglucky Thanks. We have backport this pr. It allows the broker to 
concurrently process delayed messages of different levels. But for a single 
level, its performance is not enough, only less than 1000TPS.
   
   What I mean is that we can separate the retry message from the delayed 
message, and use another customized delayed service for the retry message, so 
that even if a large number of consumption fails, it will not affect the 
delayed message. I understand that for retry messages, time inaccuracies due to 
performance are tolerable; but for delayed messages, this will have a very 
serious impact on users.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to