feature: logical cluster migration pagination

[ntnn]

Feature Description

The LogicalClusterDumps is currently dumping the entire etcd store for a logical cluster.
This poses several problems:

1. Connection problems or shard restart/failure will cause the entire dataset to be transmitted again
2. Large LCs will take a long time and block
3. No feedback on the LCMigration itself
4. Large LCs will cause considerable memory pressure

Proposed Solution

LCDumps should be paginated and/or streamed.
Will probably require changes on the LCMigration.Status to keep track of progress.

Alternative Solutions

No response

Want to contribute?

• I would like to work on this issue.

Additional Context

No response