cluster Unhealthy after replacing a unit

[macchese]

ubuntu@juju:~$ juju status etcd
Model Controller Cloud/Region Version SLA Timestamp
openstack maas-controller1 maas-cloud/default 3.2.4 unsupported 10:35:09Z

App Version Status Scale Charm Channel Rev Exposed Message
etcd 3.4.22 blocked 3 etcd stable 760 no UnHealthy with 3 known peers

Unit Workload Agent Machine Public address Ports Message
etcd/7* blocked idle 7/lxd/26 192.168.70.93 2379/tcp UnHealthy with 3 known peers
etcd/8 blocked idle 8/lxd/27 192.168.70.92 2379/tcp UnHealthy with 3 known peers
etcd/10 waiting idle 3/lxd/47 192.168.70.155 Waiting to retry etcd registration

Machine State Address Inst id Base AZ Message
3 started 192.168.6.113 op4 ubuntu@22.04 default Deployed
3/lxd/47 started 192.168.70.155 juju-206dbb-3-lxd-47 ubuntu@22.04 default Container started
7 started 192.168.6.101 xen01 ubuntu@22.04 xensrv Deployed
7/lxd/26 started 192.168.70.93 juju-206dbb-7-lxd-26 ubuntu@22.04 xensrv Container started
8 started 192.168.6.110 op1 ubuntu@22.04 default Deployed
8/lxd/27 started 192.168.70.92 juju-206dbb-8-lxd-27 ubuntu@22.04 default Container started

I removed an unit using --force (without it didn't work) and then added again one unit (etcd/10).
looking into the new unit (etcd/10) log:

ubuntu@juju:~$ juju debug-log --include etcd/10
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/etcd.py:440:register_node_with_leader
unit-etcd-10: 10:28:30 ERROR unit.etcd/10.juju-log ['/snap/bin/etcd.etcdctl', '--endpoint', 'https://192.168.70.93:2379', 'member', 'add', 'etcd10', 'https://192.168.70.155:2380']
unit-etcd-10: 10:28:30 ERROR unit.etcd/10.juju-log {'ETCDCTL_API': '2', 'ETCDCTL_CA_FILE': '/var/snap/etcd/common/ca.crt', 'ETCDCTL_CERT_FILE': '/var/snap/etcd/common/server.crt', 'ETCDCTL_KEY_FILE': '/var/snap/etcd/common/server.key'}
unit-etcd-10: 10:28:30 ERROR unit.etcd/10.juju-log b'client: etcd cluster is unavailable or misconfigured; error #0: client: etcd member https://192.168.70.92:2379 has no leader\n; error #1: client: endpoint https://192.168.70.91:2379 exceeded header timeout\n; error #2: client: etcd member https://192.168.70.93:2379 has no leader\n\n'
unit-etcd-10: 10:28:30 ERROR unit.etcd/10.juju-log None
unit-etcd-10: 10:28:30 WARNING unit.etcd/10.juju-log Notice: Unit failed self registration
unit-etcd-10: 10:28:30 INFO unit.etcd/10.juju-log etcdctl.register failed, will retry
unit-etcd-10: 10:28:30 INFO unit.etcd/10.juju-log Invoking reactive handler: hooks/relations/tls-certificates/requires.py:80:joined:certificates
unit-etcd-10: 10:28:30 INFO unit.etcd/10.juju-log status-set: waiting: Waiting to retry etcd registration
unit-etcd-10: 10:28:31 INFO juju.worker.uniter.operation ran "update-status" hook (via explicit, bespoke hook script)
unit-etcd-10: 10:28:27 INFO unit.etcd/10.juju-log Reactive main running for hook update-status
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Initializing Snap Layer
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Initializing Leadership Layer (is follower)
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/tls_client.py:18:store_ca
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/tls_client.py:44:store_server
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/tls_client.py:71:store_client
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/etcd.py:144:set_app_version
unit-etcd-10: 10:28:28 INFO unit.etcd/10.juju-log Invoking reactive handler: reactive/etcd.py:158:prepare_tls_certificates

It seems to me that the removed unit remains int the cluster (client: endpoint https://192.168.70.91:2379 exceeded header timeout) and the cluster has no leader since I removed the unit.

How to recover?