how to solve backlog quota exceeds when another subscription reset the cursor to retention position

for a topic A ， set backlog quota 1G， retention 1T， strategy is produce exception

there has sub1, which is consuming normally,

and then sub2 want to consume the retention data， then reset the cursor，then backlog size of topic will be 1G+1T which is bigger than backlog quota, producer will fail .

except for setting backlog quota -1 (which will make broker serve much data), is there any better way?

maybe wen consume retention data ， it should not be computed into backlogsize quota check？

Start by reproducing the described topic A scenario with a 1G backlog quota, 1T retention, sub1 consuming normally, and sub2 resetting its cursor to retention data. Trace how the cursor reset affects backlog size and producer behavior; done would require an agreed handling for retained data that avoids the quota failure, plus validation of the chosen behavior.

Written by the indexing model from the issue text.