CB90024 at risk of false positives on >=6.6.3

The check in CB90024 checks for accepted sync writes > DCP max buffer writes - this is accurate between 6.5.0 and 6.6.3, but since accepted sync writes will constantly increase on any cluster that's using durable writes, this will cause false positives on 6.6.3 and above once the cluster has accepted ~50,000,000 total sync writes.

We should restrict that check to running only on versions between 6.5.0 (inclusive) and 6.6.3 (exclusive).