Fix for LaunchConfigName nor LaunchTemplate existing on ASG object

[jwitko]

SUMMARY

When attempting to use ec2_asg to make changes to an already existing autoscaling group, the module fails.

Fixes #50562

ISSUE TYPE

• Bugfix Pull Request

COMPONENT NAME

ec2_asg.py

ADDITIONAL INFORMATION

I was hiting an error when running my playbooks to modify auto-scaling groups and discovered the existing issue #50562. Taking a further look I was able to get the following error message after adding a print debug line

The full traceback is:
Traceback (most recent call last):
  File "<stdin>", line 113, in <module>
  File "<stdin>", line 105, in _ansiballz_main
  File "<stdin>", line 48, in invoke_module
  File "/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py", line 1675, in <module>
  File "/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py", line 1661, in main
  File "/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py", line 1023, in create_autoscaling_group
  File "/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py", line 636, in get_properties
KeyError: 'LaunchTemplate'

failed: [localhost] (item=test-swarm-general-master-m5-large-master) => {
    "changed": false,
    "item": [
        {
            "associate_public_ip_address": false,
            "block_device_mappings": [],
            "classic_link_vpc_security_groups": [],
            "created_time": "2019-04-01 20:07:19.322000+00:00",
            "ebs_optimized": true,
            "iam_instance_profile": "test-swarm-general-role",
            "image_id": "ami-056d69b6d20f9f180",
            "instance_monitoring": false,
            "instance_type": "m5.large",
            "kernel_id": "",
            "key_name": "ansible-keypair",
            "launch_configuration_arn": "arn:aws:autoscaling:us-east-1:REDACTED:launchConfiguration:dd954e0f-8e3c-463c-b30b-41b1b698fbb0:launchConfigurationName/test-swarm-general-m5-large-master-20190401-1605",
            "launch_configuration_name": "test-swarm-general-m5-large-master-20190401-1605",
            "ramdisk_id": "",
            "security_groups": [
                "sg-02f86b76"
            ],
            "user_data": ""
        },
        {
            "alb_target_groups": [
                {
                    "alb_tg_health_check_interval": 120,
                    "alb_tg_health_check_path": "/v2/_catalog",
                    "alb_tg_health_check_port": 5000,
                    "alb_tg_health_check_protocol": "http",
                    "alb_tg_health_check_timeout": 15,
                    "alb_tg_healthy_threshold_count": 3,
                    "alb_tg_name": "us-east-1-test-dockerhub-5000",
                    "alb_tg_port": 5000,
                    "alb_tg_protocol": "http",
                    "alb_tg_return_code_matcher": 200,
                    "alb_tg_unhealthy_threshold_count": 3
                },
                {
                    "alb_tg_health_check_interval": 120,
                    "alb_tg_health_check_path": "/users/sign_in",
                    "alb_tg_health_check_port": "8085",
                    "alb_tg_health_check_protocol": "http",
                    "alb_tg_health_check_timeout": 15,
                    "alb_tg_healthy_threshold_count": 3,
                    "alb_tg_name": "test-swarm-general-gitlab",
                    "alb_tg_port": "8085",
                    "alb_tg_protocol": "http",
                    "alb_tg_return_code_matcher": 302,
                    "alb_tg_unhealthy_threshold_count": 3
                }
            ],
            "asg_alarm_arn": "arn:aws:sns:us-east-1:569557445448:loadbalancer_alarms",
            "asg_default_cooldown": 700,
            "asg_desired": 3,
            "asg_dimension": {
                "AutoScalingGroupName": "test-swarm-general-master"
            },
            "asg_evaluation_periods": 3,
            "asg_health_check_period": 300,
            "asg_max": 3,
            "asg_metric_namespace": "AWS/EC2",
            "asg_min": 3,
            "asg_period": 120,
            "asg_purpose": "master",
            "asg_scale_in_cooldown": 700,
            "asg_scale_metric": "CPUUtilization",
            "asg_scale_out_cooldown": 700,
            "asg_scalein_comparison": "<=",
            "asg_scalein_description": "This will alarm when CPU utilization is under 50% for 9 minutes",
            "asg_scalein_threshold": "30.0",
            "asg_scaleout_comparison": ">=",
            "asg_scaleout_description": "This will alarm when CPU utilization is over 80% for 9 minutes",
            "asg_scaleout_threshold": "80.0",
            "asg_statistic": "Average",
            "asg_unit": "Percent",
            "swarm_labels": [
                "swarm-general:true",
                "master:true"
            ],
            "type": "m5.large"
        }
    ],
    "module_stderr": "Traceback (most recent call last):\n  File \"<stdin>\", line 113, in <module>\n  File \"<stdin>\", line 105, in _ansiballz_main\n  File \"<stdin>\", line 48, in invoke_module\n  File \"/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py\", line 1675, in <module>\n  File \"/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py\", line 1661, in main\n  File \"/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py\", line 1023, in create_autoscaling_group\n  File \"/tmp/ansible_ec2_asg_payload_6Npffj/__main__.py\", line 636, in get_properties\nKeyError: 'LaunchTemplate'\n",
    "module_stdout": "TESTING: {u'InstanceId': 'i-007caa4fef0eeb33f', u'ProtectedFromScaleIn': False, u'AvailabilityZone': 'us-east-1d', u'HealthStatus': 'Healthy', u'LifecycleState': 'InService'}\n",
    "msg": "MODULE FAILURE\nSee stdout/stderr for the exact error",
    "rc": 1
}

The print out was from the section where the error was hitting:

        for i in autoscaling_group_instances:
            print "TESTING: " + str(i)
            if i.get('LaunchConfigurationName'):
                instance_facts[i['InstanceId']] = {'health_status': i['HealthStatus'],
                                                   'lifecycle_state': i['LifecycleState'],
                                                   'launch_config_name': i['LaunchConfigurationName']}
            else:
                instance_facts[i['InstanceId']] = {'health_status': i['HealthStatus'],
                                                   'lifecycle_state': i['LifecycleState'],
                                                   'launch_template': i['LaunchTemplate']}

It turns out as evidenced by "module_stdout": "TESTING: {u'InstanceId': 'i-007caa4fef0eeb33f', u'ProtectedFromScaleIn': False, u'AvailabilityZone': 'us-east-1d', u'HealthStatus': 'Healthy', u'LifecycleState': 'InService'}\n", that there are in fact times where this will return without either LaunchConfigurationName or LaunchTemplate defined. (The ASG in question has a LaunchConfig).

This PR adds a conditional to allow for neither to be defined and it does indeed seem to solve my issue without any unintended side-effects that I can tell.

[jwitko]

@nwsparks Would love to hear from you if the above resolves your issue.

[ansibot]

cc @garethr @ryansb @s-hertel
click here for bot help

[willthames]

Are instances in this state on the way in or the way out of the ASG? (I can only imagine an instance not to have a launch config/template in one of those two scenarios). If they're on the way out we should probably just not put them in the instance_facts

[jwitko]

@willthames This was done on an ASG that was not in any transitional state. I dug a little deeper and I believe this is because in my organization we rotate and prune launch configs at a very fast pace. The instances that are in this ASG no longer have launch configurations attached to them because those launch configurations were deleted after the instances were created.

So the scenario this covers is if you have instances that were created and then later on their launch configurations were deleted (possible the same behavior for LaunchTemplate, I'm unsure)

[jwitko]

@willthames checking if there is anything else I can provide for this PR?

[willthames]

It would be good to improve theinstance_facts docs to reflect this but they don't detail every key already so the sample is still accurate most of the time