Hello Guys,
This is the first time I write to the group, so please bare with me if I miss something or my English is bad.
I want to create ec2 instances and use the output of the ec2 module as input for other tasks, like the description in the documentation. Unfortunately, unlike the documentation I have a problem using the ec2 variable that I register. The documentation states that to get the instance ID I need to do ec2.instanses.id, but actually I had to dive deeper and manually set it to ec2.results[0].instances[0].id. This is very error prone because the underlaying json returned by the ec2 module constantly changes the location of ec2.results and/or instances.
Am I doing something wrong? I really want to use the simple format ec2.instances.id 
Here is the current example
- name: Provision NAT instance
local_action:
module: ec2
region: "{{ region }}"
key_name: "key-dev"
instance_type: "{{ ec2_nat_type }}"
image: "{{ ec2_nat_image }}"
wait: yes
group: "{{ rcrtx_env }}_sg_nat"
instance_tags:
Name: "{{ rcrtx_env }}_nat"
env: "{{ rcrtx_env }}"
component: "nat"
vpc_subnet_id: "{{ item.id }}"
source_dest_check: no
monitoring: yes
exact_count: 1
count_tag:
component: "nat"
with_items: vpc.subnets
when: item.resource_tags.Name == "{{ rcrtx_env }}_subnet_public_1a"
register: ec2nat
- name: Associate new EIP for the NAT instance
tags: eip
local_action:
module: ec2_eip
in_vpc: yes
region: "{{ region }}"
instance_id: "{{ ec2nat.results[0].instances[0].id }}"
when: ec2nat.results[0].instances[0].id is defined