Lambda function for propagating EKS managed nodegroup tags

handler.py

#!/usr/bin/env python

import boto3
import botocore.exceptions
import logging
from time import sleep

MAX_ATTEMPTS = 100

def handler(event, context):
    detail = event['detail']
    if detail['eventSource'] != 'eks.amazonaws.com' or detail['eventName'] != 'CreateNodegroup':
        print('Received event {} from {} - skipping'.format(detail['eventSource'], detail['eventName']))
        return

    cluster_name = detail['requestParameters']['name']
    nodegroup = detail['responseElements']['nodegroup']
    nodegroup_name = nodegroup['nodegroupName']
    nodegroup_tags = nodegroup['tags']

    eks = boto3.client('eks')
    autoscaling = boto3.client('autoscaling')
    ec2 = boto3.client('ec2')

    print('[{}] Waiting for nodegroup {}'.format(cluster_name, nodegroup_name))

    attempts = 0
    autoscaling_group_name = None

    while not autoscaling_group_name:
        attempts = attempts + 1
        if attempts > MAX_ATTEMPTS:
            raise TimeoutError('[{}] Timed out waiting to obtain Auto Scaling Group Name for nodegroup {}'.format(cluster_name, nodegroup_name))
        try:
            response = eks.describe_nodegroup(
                clusterName=cluster_name,
                nodegroupName=nodegroup_name
            )
            status = response['nodegroup']['status']
            if status == 'DELETING':
                print('[{}] Nodegroup {} is {}'.format(cluster_name, nodegroup_name, status))
                return
            elif status == 'CREATING':
                sleep(5)
                continue
            elif status == 'ACTIVE':
                autoscaling_group_name = response['nodegroup']['resources']['autoScalingGroups'][0]['name']
                break
            else:
                raise ValueError('[{}] Unknown nodegroup status: {} is {}'.format(cluster_name, nodegroup_name, status))
        except KeyError:
            sleep(5)
        except botocore.exceptions.ClientError as e:
            if e.response['Error']['Code'] == 'ResourceNotFoundException':
                sleep(5)
            else:
                raise

    print('[{}] Tagging Auto Scaling Group {}'.format(cluster_name, autoscaling_group_name))

    autoscaling.create_or_update_tags(Tags=[
        {
            'ResourceId': autoscaling_group_name,
            'ResourceType': 'auto-scaling-group',
            'Key': tag_key,
            'Value': tag_value,
            'PropagateAtLaunch': True,
        }
    for (tag_key, tag_value) in nodegroup_tags.items() if not tag_key.startswith('aws:')])

    print('[{}] Tagging running instances in Auto Scaling Group {}'.format(cluster_name, autoscaling_group_name))

    response = autoscaling.describe_auto_scaling_groups(
        AutoScalingGroupNames=[autoscaling_group_name]
    )
    instance_ids = [i['InstanceId'] for i in response['AutoScalingGroups'][0]['Instances']]
    ec2.create_tags(
        Resources=instance_ids,
        Tags=[{
            'Key': tag_key,
            'Value': tag_value,
        } for (tag_key, tag_value) in nodegroup_tags.items() if not tag_key.startswith('aws:')]
    )