percona-cluster charm errors on "cluster-relation-departed" when tearing down a model

e.g.

Model Controller Cloud/Region Version SLA Timestamp Notes
zaza-eb5f8cc00b42 auto-osci-sv03 serverstack/serverstack 2.8.7 unsupported 14:06:46Z attempt 2 to destroy model failed (will retry): model not empty, found 3 machines, 3 applications (model not empty)

App Version Status Scale Charm Store Rev OS Notes
hacluster maintenance 3 hacluster jujucharms 160 ubuntu
keystone waiting 0 keystone jujucharms 528 ubuntu
percona-cluster 5.7.20 error 3 percona-cluster local 45 ubuntu

Unit Workload Agent Machine Public address Ports Message
percona-cluster/0* error idle 1 172.17.103.20 3306/tcp hook failed: "cluster-relation-departed"
  hacluster/2* maintenance executing 172.17.103.20 (stop) stopping charm software
percona-cluster/1 blocked executing 2 172.17.103.21 3306/tcp MySQL is down. Sequence Number: -1. Safe To Bootstrap: 0
  hacluster/1 maintenance executing 172.17.103.21 (stop) stopping charm software
percona-cluster/2 blocked executing 3 172.17.103.19 3306/tcp Insufficient peers to bootstrap cluster
  hacluster/0 maintenance executing 172.17.103.19 (stop) stopping charm software

Machine State DNS Inst id Series AZ Message
1 stopped 172.17.103.20 099a4e91-1b45-4a6f-97aa-1c8bb1eca12b bionic nova ACTIVE
2 stopped 172.17.103.21 b46d599c-cfc4-4d28-bac4-7ec40897f493 bionic nova ACTIVE
3 stopped 172.17.103.19 3309e7ee-7f4f-4fdc-a2cc-570eea9f4331 bionic nova ACTIVE