several hundreds of 4TB HGST enterprise in a cluster at university. Disks get replaced ASAP but I'm seeing the checksum errors often in the logfiles when a disk

several hundreds of 4TB HGST enterprise in a cluster at university. Disks get replaced ASAP but I'm seeing the checksum errors often in the logfiles when a disk starts to collapse. Btw: So far the Backblaze reliability numbers check out (<1% annual failure rate)