juju wait-for unit checks application message

[Marian Gasparovic]

When I shutdown one unit of an application and then try to check wait-for message it shows wrong information

$ juju status aodh/1
Model Controller Cloud/Region Version SLA Timestamp
openstack foundations-maas maas_cloud/default 2.9.37 unsupported 16:03:11Z

App Version Status Scale Charm Channel Rev Exposed Message
aodh 14.0.0 active 0/1 aodh yoga/stable 77 no Unit is ready
aodh-mysql-router 8.0.31 active 0/1 mysql-router 8.0/stable 35 no Unit is ready
hacluster-aodh active 0/1 hacluster 2.4/stable 117 no Unit is ready and clustered
logrotated active 0/1 logrotated stable 7 no Unit is ready.
public-policy-routing active 0/1 advanced-routing stable 11 no Unit is ready

Unit Workload Agent Machine Public address Ports Message
aodh/1 unknown lost 1/lxd/0 10.246.166.156 8042/tcp agent lost, see 'juju show-status-log aodh/1'
  aodh-mysql-router/0 unknown lost 10.246.166.156 agent lost, see 'juju show-status-log aodh-mysql-router/0'
  hacluster-aodh/0 unknown lost 10.246.166.156 agent lost, see 'juju show-status-log hacluster-aodh/0'
  logrotated/37 unknown lost 10.246.166.156 agent lost, see 'juju show-status-log logrotated/37'
  public-policy-routing/24 unknown lost 10.246.166.156 agent lost, see 'juju show-status-log public-policy-routing/24'

$ juju wait-for unit aodh/1
Unit "aodh/1" is running
properties:
  life: alive
  workload-status: active

It looks like it is getting status and message from the application, not from the unit

$ juju wait-for unit aodh/1 --query 'workload-message=="agent lost" || workload-status == "active"'
Unit "aodh/1" is running
properties:
  workload-message: Unit is ready
  workload-status: active