Test versus production

The current cluster is a test bed — three nodes with a 150 GiB Longhorn data disk each. The production cluster is sized far larger, so the disk-pressure and per-node footprint concerns of running two HA engines plus the LGTM stack do not carry over. The design above is written for production; on the test cluster, size PVCs down accordingly.