platinumhamburg opened a new issue, #1376:
URL: https://github.com/apache/fluss/issues/1376

   ### Search before asking
   
   - [x] I searched in the [issues](https://github.com/alibaba/fluss/issues) 
and found nothing similar.
   
   
   ### Motivation
   
   The CoordinatorEventManager currently uses a single-threaded model to 
process coordination events. In large-scale clusters, DDL events or even an 
increasing bucket count can generate thousands of pending events, resulting in 
event processing times stretching to minutes. However, there is no 
transactional ordering requirement between all different types of events, 
meaning strict serialized processing is unnecessary.
   
   We can categorize all event types into several event families:
   - DDL Events
   - Cluster Events
   - Commit Events
   - Others
   Only DDL events and cluster events require strict ordering within their 
respective groups.
   
   So, there are multiple optimization opportunities to explore based on this 
grouping and relaxed ordering requirements.
   
   ### Solution
   
   - [ ] CoordinatorEventManager support multiple EventProcessor for concurrent 
event processing
   - [ ] CoordinatorContext support multi-threaded access
   - [ ] Remove AccessContextEvent from Event Process Loop
   - [ ] Use RemoteFileDeletionQueue instead of processing 
DeleteReplicaResponseReceivedEvent in EventProcessor
   - [ ] Add rack lable for TabletServerMetricGroup
   - [ ] Add  EventType lable for all event related metrics 
   
   ### Anything else?
   
   _No response_
   
   ### Willingness to contribute
   
   - [ ] I'm willing to submit a PR!


-- 
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