FrankChen021 commented on PR #18136: URL: https://github.com/apache/druid/pull/18136#issuecomment-2968760242
> @FrankChen021 thanks for the PR! Curious, how is this different from [broker tiering](https://druid.apache.org/docs/latest/operations/mixed-workloads/#broker-tiering)? Wouldn’t it be possible to achieve the same kind of preferred tier selection across AZs by configuring tiers on the brokers? 1. broker tiering requires the router to route the queries to different brokers, which introduces extra deployment/dependency. Since router is an optional component, there should be a way that the broker can handle routing. 2. broker tiering requires the client the set some query context parameters for routing. in another word, It's the client that is responsible for routing, which is not the case here 3. broker tiering does not provide a way to achieve fallback routing. If historical nodes in one AZ are down, we need to route queries to historicals nodes in another AZ -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
