如何查询 TKE 节点列表¶

from tencentcloud.common import credential
from tencentcloud.tke.v20180525 import tke_client, models

# 初始化认证
cred = credential.Credential("SecretId", "SecretKey")
client = tke_client.TkeClient(cred, "ap-guangzhou")

# 方式1: 查询所有节点
req = models.DescribeClusterInstancesRequest()
req.ClusterId = "cls-xxxxxxxx"

resp = client.DescribeClusterInstances(req)

print(f"节点总数: {resp.TotalCount}")
for instance in resp.InstancesList:
    print(f"节点ID: {instance.InstanceId}")
    print(f"节点名称: {instance.InstanceName}")
    print(f"节点状态: {instance.InstanceState}")
    print(f"内网IP: {instance.LanIP}")
    print(f"节点类型: {instance.InstanceType}")
    print(f"可用区: {instance.Zone}")
    print("-" * 40)

# 方式2: 查询指定节点
req = models.DescribeClusterInstancesRequest()
req.ClusterId = "cls-xxxxxxxx"
req.InstanceIds = ["ins-xxxxxxxx"]

resp = client.DescribeClusterInstances(req)

if resp.TotalCount > 0:
    instance = resp.InstancesList[0]
    print(f"节点详情:")
    print(f"  ID: {instance.InstanceId}")
    print(f"  名称: {instance.InstanceName}")
    print(f"  状态: {instance.InstanceState}")
    print(f"  CPU: {instance.CPU} 核")
    print(f"  内存: {instance.Mem} GB")
    print(f"  操作系统: {instance.OsName}")

# 方式3: 使用过滤条件
req = models.DescribeClusterInstancesRequest()
req.ClusterId = "cls-xxxxxxxx"
req.Filters = [
    models.Filter(Name="NodePoolId", Values=["np-xxxxxxxx"]),
    models.Filter(Name="InstanceState", Values=["running"])
]

resp = client.DescribeClusterInstances(req)
print(f"找到 {resp.TotalCount} 个运行中的节点（节点池 np-xxxxxxxx）")

# 方式4: 分页查询
def get_all_nodes(client, cluster_id):
    """获取所有节点（分页）"""
    all_instances = []
    offset = 0
    limit = 100

    while True:
        req = models.DescribeClusterInstancesRequest()
        req.ClusterId = cluster_id
        req.Limit = limit
        req.Offset = offset

        resp = client.DescribeClusterInstances(req)
        all_instances.extend(resp.InstancesList)

        if len(resp.InstancesList) < limit:
            break

        offset += limit

    return all_instances

all_nodes = get_all_nodes(client, "cls-xxxxxxxx")
print(f"总共 {len(all_nodes)} 个节点")

# 验证响应
assert resp.TotalCount > 0, "集群中没有节点"
assert len(resp.InstancesList) > 0, "节点列表为空"

# 验证节点信息完整
instance = resp.InstancesList[0]
assert instance.InstanceId is not None, "节点 ID 为空"
assert instance.InstanceState in ["running", "initializing", "abnormal"], "节点状态异常"

def get_cluster_resources(client, cluster_id):
    """统计集群节点资源"""
    req = models.DescribeClusterInstancesRequest()
    req.ClusterId = cluster_id

    resp = client.DescribeClusterInstances(req)

    total_cpu = 0
    total_mem = 0
    node_count = resp.TotalCount

    for instance in resp.InstancesList:
        if instance.InstanceState == "running":
            total_cpu += instance.CPU
            total_mem += instance.Mem

    print(f"集群资源统计:")
    print(f"  节点数量: {node_count}")
    print(f"  总 CPU: {total_cpu} 核")
    print(f"  总内存: {total_mem} GB")

    return {
        "node_count": node_count,
        "total_cpu": total_cpu,
        "total_mem": total_mem
    }

# 使用示例
resources = get_cluster_resources(client, "cls-xxxxxxxx")

from collections import defaultdict

def group_nodes_by_nodepool(client, cluster_id):
    """按节点池分组节点"""
    req = models.DescribeClusterInstancesRequest()
    req.ClusterId = cluster_id

    resp = client.DescribeClusterInstances(req)

    nodepool_groups = defaultdict(list)

    for instance in resp.InstancesList:
        nodepool_id = instance.NodePoolId or "default"
        nodepool_groups[nodepool_id].append(instance)

    print("节点池分组:")
    for nodepool_id, instances in nodepool_groups.items():
        print(f"  {nodepool_id}: {len(instances)} 个节点")
        for inst in instances:
            print(f"    - {inst.InstanceId} ({inst.InstanceState})")

    return nodepool_groups

# 使用示例
groups = group_nodes_by_nodepool(client, "cls-xxxxxxxx")

def detect_abnormal_nodes(client, cluster_id):
    """检测异常节点"""
    req = models.DescribeClusterInstancesRequest()
    req.ClusterId = cluster_id

    resp = client.DescribeClusterInstances(req)

    abnormal_nodes = []

    for instance in resp.InstancesList:
        if instance.InstanceState in ["abnormal", "failed"]:
            abnormal_nodes.append({
                "instance_id": instance.InstanceId,
                "instance_name": instance.InstanceName,
                "state": instance.InstanceState,
                "failed_reason": instance.FailedReason or "未知"
            })

    if abnormal_nodes:
        print(f"发现 {len(abnormal_nodes)} 个异常节点:")
        for node in abnormal_nodes:
            print(f"  - {node['instance_id']} ({node['state']}): {node['failed_reason']}")
    else:
        print("所有节点状态正常")

    return abnormal_nodes

# 使用示例
abnormal = detect_abnormal_nodes(client, "cls-xxxxxxxx")

import csv

def export_nodes_to_csv(client, cluster_id, filename="nodes.csv"):
    """导出节点信息到 CSV"""
    req = models.DescribeClusterInstancesRequest()
    req.ClusterId = cluster_id

    resp = client.DescribeClusterInstances(req)

    with open(filename, 'w', newline='', encoding='utf-8') as f:
        writer = csv.writer(f)
        writer.writerow([
            "节点ID", "节点名称", "状态", "内网IP", "可用区",
            "机型", "CPU", "内存", "操作系统", "节点池ID", "创建时间"
        ])

        for instance in resp.InstancesList:
            writer.writerow([
                instance.InstanceId,
                instance.InstanceName,
                instance.InstanceState,
                instance.LanIP,
                instance.Zone,
                instance.InstanceType,
                instance.CPU,
                instance.Mem,
                instance.OsName,
                instance.NodePoolId or "-",
                instance.CreatedTime
            ])

    print(f"已导出 {resp.TotalCount} 个节点信息到 {filename}")

# 使用示例
export_nodes_to_csv(client, "cls-xxxxxxxx")