Role fails after service start

Summary

When rocketchat service starts again (and is not restart task but after upgrgdaing to service role from 1.4.0 → 1.5.0 ) role do not wait for application to become ready.

Steps to reproduce

1. execute deploy playbook with service role - 1.4.0
2. then execute the same playbook with service role - 1.5.0

Expected behavior

Playbook ends with success.

Actual behavior

Playbook fails on login into Rocket.Chat task

Relevant logs and/or screenshots

TASK [service : add docker-rocketchat service] ************************************************************************
Thursday 18 April 2019  09:46:57 +0000 (0:00:00.076)       0:02:24.738 ******** 
changed: [rocketchat]

TASK [service : start the service] ************************************************************************************
Thursday 18 April 2019  09:46:58 +0000 (0:00:00.880)       0:02:25.619 ******** 
changed: [rocketchat]

TASK [rocketchat : restart service if necessary] **********************************************************************
Thursday 18 April 2019  09:47:27 +0000 (0:00:28.835)       0:02:54.455 ******** 
skipping: [rocketchat]

TASK [rocketchat : login into Rocket.Chat] ****************************************************************************
Thursday 18 April 2019  09:47:27 +0000 (0:00:00.058)       0:02:54.514 ******** 
fatal: [rocketchat]: FAILED! => {"changed": false, "content": "", "failed_when_result": true, "msg": "Status code was -1 and not [200]: Request failed: <urlopen error [Errno 111] Connection refused>", "redirected": false, "status": -1, "url": "http://127.0.0.1:80/api/v1/login"}

Possible fixes

Move service_restart tasks to deploy.yml file and only restart service if necessary task tag with when 'flag'