How to delete cloudwatch alarms that are attached to non-existent instances? - amazon-ec2

Question

When I create new EC2 instances I use an ansible dynamic inventory to create new cloudwatch metrics alarms. So far so good:
- name: set AWS CloudWatch alarms
hosts: tag_env_production
vars:
alarm_slack: 'arn:aws:sns:123:metrics-alarms-slack'
tasks:
- name: "CPU > 70%"
ec2_metric_alarm:
state: present
name: "{{ ec2_tag_Name }}-CPU"
region: "{{ ec2_region }}"
dimensions:
InstanceId: '{{ ec2_id }}'
namespace: "AWS/EC2"
metric: CPUUtilization
statistic: Average
comparison: ">="
threshold: 70.0
unit: Percent
period: 300
evaluation_periods: 1
description: Triggered when CPU utilization is more than 70% for 5 minutes
alarm_actions: ['{{ alarm_slack }}']
when: ec2_tag_group == 'lazyservers'
Executing as follows:
ansible-playbook -v ec2_alarms.yml -i inventories/ec2/ec2.py
After creating the new instances I drop the old ones (manually). The problem is that I'd need to delete the alarms for existing metrics attached to the old instances.
Am I missing something or there is no way how to do this via the dynamic inventory?
My current idea is to delete the metrics for instances that are in the "Terminating" state, but the downside is that if I run the playbook after those instances are terminated, they simply won't be visible.

Answer 1

Before delete the instance, delete the alarm try with something like this:
- name: delete alarm
ec2_metric_alarm:
state: absent
region: ap-southeast-2
name: "cpu-low"
metric: "CPUUtilization"
namespace: "AWS/EC2"
statistic: Average
comparison: "<="
threshold: 5.0
period: 300
evaluation_periods: 3
unit: "Percent"
description: "This will alarm when a bamboo slave's cpu usage average is lower than 5% for 15 minutes "
dimensions: {'InstanceId':'{{ instance_id }}'}
alarm_actions: ["action1","action2"]