Ansible determines success of a playbook by the success or failure of the individual tasks in the play. Sometimes this method is limiting because it does not account for higher level logic that might dictate a failure however, there is a module that can be used to account for this. The fail module allows Ansible to process a conditional statement and will fail the play if the condition fails. Lets look at an example.

---
- hosts: all
  gather_facts: no

  tasks:
  - name: check for port 80
    wait_for:
      port: 80
      timeout: 5
    ignore_errors: yes
    register: check_http

  - name: check for port 443
    wait_for:
      port: 443
      timeout: 5
    ignore_errors: yes
    register: check_https

  - fail:
      msg: "Port 80 and 443 are not open and listening"
    when: check_https.failed and check_http.failed

In this example, we have have two tasks checking if a port 80 and 443 are open on a target host. We only want to fail if both ports are not open. To do this, we add ignore_errors to to each of our wait_for tasks. See the ignore_errors post for more information on this parameter. This will prevent Ansible from failing on that task. We also need to register the output of each wait_for task for later. After both of our tests, we can use the fail module to evaluate if the play is successful or not. The msg parameter allows us to log a custom message when this module runs and a simple when condition checks for our condition. See the complete documentation here.