During chunk migration, we run a snapshot read concern aggregation on the config server to get the current global indexes. If the config server node that gets contacted has recently restarted, we may not be able to get a Snapshot above the minVisible timestamp. This can cause the migration to fail with SnapshotUnavailable.