Currently, when we externally set the storage radius with the depthmonitor, the change does not trigger an eviction process so the reserve size value does not accurately reflect the number of chunks with POs >= storage radius.
In the current implementation when an eviction is deemed necessary by the localstore, the localstore calls the batchstore callback to figure out which batches to evict and up to what storage radius.
A potential fix here is that on top of the localstore -> batchstore callback, we need a new callback to call the localstore when storage radius changes.
Currently, when we externally set the storage radius with the depthmonitor, the change does not trigger an eviction process so the reserve size value does not accurately reflect the number of chunks with POs >= storage radius.
In the current implementation when an eviction is deemed necessary by the localstore, the localstore calls the batchstore callback to figure out which batches to evict and up to what storage radius.
A potential fix here is that on top of the localstore -> batchstore callback, we need a new callback to call the localstore when storage radius changes.