Using serial strategy for playbook imported in a toplevel playbook causes entire play termination if some host is unreachable

Hello all.
Here is a simplified example:

`

  • import_playbook: test_unavailable_node_different_strategy.yml

  • hosts: 127.0.0.1
    connection: local

The question is would we end up here in case of imported playbook (above) touches unavailable host

tasks:

  • debug: msg="Playbook finished"Введите код…
    `

And the code for imported playbook:

`

  • hosts: all

Comment the line below and everything would be fine

serial: 1

tasks:

  • debug: msg="{{ inventory_hostname }}"Введите код…
    `

So when I’m launching something like this
ansible-playbook -i '127.0.0.1,10.0.0.1' use_unavailable_nodes.yml --ask-pass -u andreyВведите код...
when the 10.0.0.1 host is unreachable, the entire play aborts before reaching the task to be performed locally (see main playbook).

If you comment ‘serial: 1’ then the local action in the main playbook also would be performed.
Why there is such behaviour? Is this a bug or I’m missing something?

I ran a playbook on batches and one host failed but just that batch stopped.