mik-laj commented on a change in pull request #16571: URL: https://github.com/apache/airflow/pull/16571#discussion_r660151177
########## File path: airflow/providers/amazon/aws/operators/eks.py ########## @@ -0,0 +1,737 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# pylint: disable=invalid-name +"""This module contains Amazon EKS operators.""" +import json +import os +from datetime import datetime +from time import sleep +from typing import Dict, List, Optional + +from boto3 import Session + +from airflow.models import BaseOperator +from airflow.providers.amazon.aws.hooks.eks import DEFAULT_PAGINATION_TOKEN, DEFAULT_RESULTS_PER_PAGE, EKSHook +from airflow.providers.amazon.aws.utils.eks_kube_config import ( + DEFAULT_CONTEXT_NAME, + DEFAULT_KUBE_CONFIG_PATH, + DEFAULT_NAMESPACE_NAME, + DEFAULT_POD_USERNAME, + generate_config_file, +) +from airflow.providers.cncf.kubernetes.operators.kubernetes_pod import KubernetesPodOperator + +CHECK_INTERVAL_SECONDS = 15 +TIMEOUT_SECONDS = 25 * 60 +CONN_ID = "eks" +REGION = Session().region_name +DEFAULT_COMPUTE_TYPE = 'nodegroup' +DEFAULT_NODEGROUP_NAME_SUFFIX = '-nodegroup' +DEFAULT_POD_NAME = 'pod' +KUBE_CONFIG_ENV_VAR = 'KUBECONFIG' + + +class EKSCreateClusterOperator(BaseOperator): + """ + Creates an Amazon EKS Cluster control plane. + + Optionally, can also create the supporting compute architecture: + If argument 'compute' is provided with a value of 'nodegroup', will also attempt to create an Amazon + EKS Managed Nodegroup for the cluster. See EKSCreateNodegroupOperator documentation for requirements. + + .. seealso:: + For more information on how to use this operator, take a look at the guide: + :ref:`howto/operator:EKSCreateClusterOperator` + + :param cluster_name: The unique name to give to your Amazon EKS Cluster. + :type cluster_name: str + :param cluster_role_arn: The Amazon Resource Name (ARN) of the IAM role that provides permissions for the + Kubernetes control plane to make calls to AWS API operations on your behalf. + :type cluster_role_arn: str + :param resources_vpc_config: The VPC configuration used by the cluster control plane. + :type resources_vpc_config: Dict + :param compute: The type of compute architecture to generate along with the cluster. + Defaults to 'nodegroup' to generate an EKS Managed Nodegroup. + :type compute: str + :param aws_conn_id: The Airflow connection used for AWS credentials. + If this is None or empty then the default boto3 behaviour is used. If + running Airflow in a distributed manner and aws_conn_id is None or + empty, then the default boto3 configuration would be used (and must be + maintained on each worker node). + :type aws_conn_id: str + + If 'compute' is 'nodegroup', the following are required: + + :param nodegroup_name: The unique name to give your EKS Managed Nodegroup. + :type nodegroup_name: str + :param nodegroup_role_arn: The Amazon Resource Name (ARN) of the IAM role to associate + with the EKS Managed Nodegroup. + :type nodegroup_role_arn: str + + """ + + def __init__( + self, + cluster_name: str, + cluster_role_arn: str, + resources_vpc_config: Dict, + nodegroup_name: Optional[str] = None, + nodegroup_role_arn: Optional[str] = None, + compute: Optional[str] = DEFAULT_COMPUTE_TYPE, + conn_id: Optional[str] = CONN_ID, + region: Optional[str] = REGION, + **kwargs, + ) -> None: + super().__init__(**kwargs) + self.clusterName = cluster_name + self.clusterRoleArn = cluster_role_arn + self.resourcesVpcConfig = resources_vpc_config + self.compute = compute + self.conn_id = conn_id + self.region = region + + if self.compute == 'nodegroup': + self.nodegroupName = nodegroup_name or self.clusterName + DEFAULT_NODEGROUP_NAME_SUFFIX + if nodegroup_role_arn: + self.nodegroupRoleArn = nodegroup_role_arn + else: + message = "Creating an EKS Managed Nodegroup requires nodegroup_role_arn to be passed in." + self.log.error(message) + raise AttributeError(message) + + def execute(self, context): + eks_hook = EKSHook( + aws_conn_id=self.conn_id, + region_name=self.region, + ) + + eks_hook.create_cluster( + name=self.clusterName, + roleArn=self.clusterRoleArn, + resourcesVpcConfig=self.resourcesVpcConfig, + ) + + if self.compute is not None: + self.log.info("Waiting for EKS Cluster to provision. This will take some time.") + + countdown = TIMEOUT_SECONDS + while eks_hook.get_cluster_state(clusterName=self.clusterName) != "ACTIVE": + if countdown >= CHECK_INTERVAL_SECONDS: + countdown -= CHECK_INTERVAL_SECONDS + self.log.info( + "Waiting for cluster to start. Checking again in %d seconds", CHECK_INTERVAL_SECONDS + ) + sleep(CHECK_INTERVAL_SECONDS) + else: + message = "Cluster is still inactive after the allocated time limit. Aborting." + self.log.error(message) + raise RuntimeError(message) + + if self.compute == 'nodegroup': + eks_hook.create_nodegroup( + clusterName=self.clusterName, + nodegroupName=self.nodegroupName, + subnets=self.resourcesVpcConfig.get('subnetIds'), + nodeRole=self.nodegroupRoleArn, + ) + + +class EKSCreateNodegroupOperator(BaseOperator): + """ + Creates am Amazon EKS Managed Nodegroup for an existing Amazon EKS Cluster. + + .. seealso:: + For more information on how to use this operator, take a look at the guide: + :ref:`howto/operator:EKSCreateNodegroupOperator` + + :param cluster_name: The name of the Amazon EKS Cluster to create the managed nodegroup in. + :type cluster_name: str + :param nodegroup_name: The unique name to give your managed nodegroup. + :type nodegroup_name: str + :param nodegroup_subnets: + The subnets to use for the Auto Scaling group that is created for the managed nodegroup. + :type nodegroup_subnets: List[str] + :param nodegroup_role_arn: + The Amazon Resource Name (ARN) of the IAM role to associate with the managed nodegroup. + :type nodegroup_role_arn: str + :param aws_conn_id: The Airflow connection used for AWS credentials. + If this is None or empty then the default boto3 behaviour is used. If + running Airflow in a distributed manner and aws_conn_id is None or + empty, then the default boto3 configuration would be used (and must be + maintained on each worker node). + :type aws_conn_id: str + + """ + + def __init__( + self, + cluster_name: str, + nodegroup_subnets: List[str], + nodegroup_role_arn: str, + nodegroup_name: Optional[str], + conn_id: Optional[str] = CONN_ID, + region: Optional[str] = REGION, + **kwargs, + ) -> None: + super().__init__(**kwargs) + self.clusterName = cluster_name + self.nodegroupSubnets = nodegroup_subnets + self.nodegroupRoleArn = nodegroup_role_arn + self.nodegroupName = nodegroup_name or cluster_name + datetime.now().strftime("%Y%m%d_%H%M%S") + self.conn_id = conn_id + self.region = region + + def execute(self, context): + eks_hook = EKSHook( + aws_conn_id=self.conn_id, + region_name=self.region, + ) + + return eks_hook.create_nodegroup( + clusterName=self.clusterName, + nodegroupName=self.nodegroupName, + subnets=self.nodegroupSubnets, + nodeRole=self.nodegroupRoleArn, + ) + + +class EKSDeleteClusterOperator(BaseOperator): + """ + Deletes the Amazon EKS Cluster control plane and all nodegroups attached to it. + + .. seealso:: + For more information on how to use this operator, take a look at the guide: + :ref:`howto/operator:EKSDeleteClusterOperator` + + :param cluster_name: The name of the Amazon EKS Cluster to delete. + :type cluster_name: str + :param aws_conn_id: The Airflow connection used for AWS credentials. + If this is None or empty then the default boto3 behaviour is used. If + running Airflow in a distributed manner and aws_conn_id is None or + empty, then the default boto3 configuration would be used (and must be + maintained on each worker node). + :type aws_conn_id: str + + """ + + def __init__( + self, cluster_name: str, conn_id: Optional[str] = CONN_ID, region: Optional[str] = REGION, **kwargs + ) -> None: + super().__init__(**kwargs) + self.clusterName = cluster_name + self.conn_id = conn_id + self.region = region + + def execute(self, context): + eks_hook = EKSHook( + aws_conn_id=self.conn_id, + region_name=self.region, + ) + + nodegroups = eks_hook.list_nodegroups(clusterName=self.clusterName).get('nodegroups') + nodegroup_count = len(nodegroups) + if nodegroup_count > 0: + self.log.info( + "A cluster can not be deleted with attached nodegroups. Deleting %d nodegroups.", + nodegroup_count, + ) + for group in nodegroups: + eks_hook.delete_nodegroup(clusterName=self.clusterName, nodegroupName=group) + + # Scaling up the timeout based on the number of nodegroups that are being processed. + additional_seconds = 5 * 60 + countdown = TIMEOUT_SECONDS + (nodegroup_count * additional_seconds) + while len(eks_hook.list_nodegroups(clusterName=self.clusterName).get('nodegroups')) > 0: + if countdown >= CHECK_INTERVAL_SECONDS: + countdown -= CHECK_INTERVAL_SECONDS + sleep(CHECK_INTERVAL_SECONDS) + self.log.info( + "Waiting for the remaining %s nodegroups to delete. Checking again in %d seconds.", + nodegroup_count, + CHECK_INTERVAL_SECONDS, + ) + else: + message = "Nodegroups are still inactive after the allocated time limit. Aborting." + self.log.error(message) + raise RuntimeError(message) + + self.log.info("No nodegroups remain, deleting cluster.") + return eks_hook.delete_cluster(name=self.clusterName) + + +class EKSDeleteNodegroupOperator(BaseOperator): + """ + Deletes an Amazon EKS Nodegroup from an Amazon EKS Cluster. + + .. seealso:: + For more information on how to use this operator, take a look at the guide: + :ref:`howto/operator:EKSDeleteNodegroupOperator` + + :param cluster_name: The name of the Amazon EKS Cluster that is associated with your nodegroup. + :type cluster_name: str + :param nodegroup_name: The name of the nodegroup to delete. + :type nodegroup_name: str + :param aws_conn_id: The Airflow connection used for AWS credentials. + If this is None or empty then the default boto3 behaviour is used. If + running Airflow in a distributed manner and aws_conn_id is None or + empty, then the default boto3 configuration would be used (and must be + maintained on each worker node). + :type aws_conn_id: str + + """ + + def __init__( + self, + cluster_name: str, + nodegroup_name: str, + conn_id: Optional[str] = CONN_ID, + region: Optional[str] = REGION, + **kwargs, + ) -> None: + super().__init__(**kwargs) + self.clusterName = cluster_name + self.nodegroupName = nodegroup_name + self.conn_id = conn_id + self.region = region + + def execute(self, context): + eks_hook = EKSHook( + aws_conn_id=self.conn_id, + region_name=self.region, + ) + + return eks_hook.delete_nodegroup(clusterName=self.clusterName, nodegroupName=self.nodegroupName) + + +class EKSDescribeAllClustersOperator(BaseOperator): + """ + Describes all Amazon EKS Clusters in your AWS account. + + :param max_results: The maximum number of results to return. + :type max_results: int + :param next_token: The nextToken value returned from a previous paginated execution. Review comment: When defining a DAG, the user should define all tasks. Here, however, there is a case where the number of tasks is not known until the first page is downloaded, and therefore cannot be defined in Airflow. In the case of Google, we've always fetched all items from all pages. This way, the user could access all elements and did not have to define dynamic DAGs. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@airflow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
