hangc0276 commented on pull request #13316: URL: https://github.com/apache/pulsar/pull/13316#issuecomment-1008585078
> only a detail... which probably fit better in a new issue > > one may even think about a config type where content could be the same for every cluster, to avoid the need to have exclusive/different configs for every single clusters > > e.g. something like a "matrix" > > Latency Matrix: > > Cluster A Cluster B Cluster C > Cluster A 0 200 900 > Cluster B 200 0 600 > Cluster C 900 600 0 @hpvd we just consider primary cluster and secondary clusters. When the primary cluster failed, we will consider the latency from the primary cluster to all secondary clusters and choose the target one according to configured policy. we doesn't need to configure latency metrics between secondary clusters. For the latency, we can get it when probe and switching. If users just want to configure the priority among secondary clusters, they can use the default `ORDER` policy, which will check the secondary cluster from the beginning to the end. -- 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]
