from kubernetes import client, config import json import re from typing import Dict, Any, List, Tuple import yaml import time def get_node_info(kubeconfig): try: # 加载配置 kubeconfig = yaml.safe_load(kubeconfig) config.load_kube_config_from_dict(kubeconfig) v1 = client.CoreV1Api() api_client = client.ApiClient() # 获取节点指标和 Pod 列表 node_metrics_path = "/apis/metrics.k8s.io/v1beta1/nodes" node_metrics_response = api_client.call_api( node_metrics_path, 'GET', auth_settings=['BearerToken'], response_type='object')[0] node_metrics = {node['metadata']['name']: node.get('usage', {}) for node in node_metrics_response.get('items', [])} # 获取所有 Pod 及其资源请求 pods = v1.list_pod_for_all_namespaces().items node_pod_resources = {} # 存储每个节点上 Pod 的资源请求 print(pods) for pod in pods: if pod.spec.node_name and pod.status.phase in ["Running", "Pending"]: node_name = pod.spec.node_name if node_name not in node_pod_resources: node_pod_resources[node_name] = { "cpu": 0, "memory": 0, "gpu": 0 } # 累加容器请求的资源 for container in pod.spec.containers: if container.resources and container.resources.requests: # CPU (转换为 millicores) cpu_request = container.resources.requests.get("cpu", "0m") cpu_millis = int(float(cpu_request.rstrip("m"))) if "m" in cpu_request else int(float(cpu_request) * 1000) node_pod_resources[node_name]["cpu"] += cpu_millis # Memory (转换为 bytes) memory_request = container.resources.requests.get("memory", "0") memory_bytes = int(float(memory_request.rstrip("KiMiGi"))) if "Ki" in memory_request: memory_bytes *= 1024 elif "Mi" in memory_request: memory_bytes *= 1024 * 1024 elif "Gi" in memory_request: memory_bytes *= 1024 * 1024 * 1024 node_pod_resources[node_name]["memory"] += memory_bytes # GPU gpu_request = container.resources.requests.get("nvidia.com/gpu", "0") node_pod_resources[node_name]["gpu"] += int(gpu_request) # 获取节点列表并计算资源使用情况 nodes = v1.list_node().items rows = [] for node in nodes: node_name = node.metadata.name internal_ip = next((address.address for address in node.status.addresses if address.type == "InternalIP"), "未分配") external_ip = next((address.address for address in node.status.addresses if address.type == "ExternalIP"), "未分配") status = node.status.conditions[-1].status if node.status.conditions else "Unknown" status = "已就绪" if status == "True" else "未就绪" # 节点角色 roles = [] role_labels = [ "node-role.kubernetes.io/control-plane", "node-role.kubernetes.io/master", "node-role.kubernetes.io/worker" ] for label in role_labels: if label in node.metadata.labels: roles.append(label.split("/")[-1]) roles_str = "控制节点" if roles else "工作节点" # 节点运行时间 running_time = time.time() - node.metadata.creation_timestamp.timestamp() node_age = running_time # 节点信息 k8s_version = node.status.node_info.kubelet_version os_image = node.status.node_info.os_image kernel_version = node.status.node_info.kernel_version container_runtime = node.status.node_info.container_runtime_version # 自定义标签 labels = node.metadata.labels kyy_labels = [f"{k}={v}" for k, v in labels.items() if k.startswith('kyy-')] # 实时资源使用情况 cpu_usage = node_metrics.get(node_name, {}).get('cpu', 'undefined') if cpu_usage and isinstance(cpu_usage, str): cpu_usage = int(cpu_usage.replace("n", "")) cpu_usage = f'{(cpu_usage / 1000000 / 1000):.3f}核' memory_usage = node_metrics.get(node_name, {}).get('memory', 'undefined') if memory_usage and isinstance(memory_usage, str): memory_usage = int(memory_usage.replace("Ki", "")) memory_usage = f"{(memory_usage / 1024 / 1024):.3f}Gi" # 节点总资源 total_cpu = float(node.status.allocatable.get("cpu", "0")) total_memory = parse_resource_value(node.status.allocatable.get("memory", "0")) / (1024 ** 1) #内存默认Mi转成Gi total_gpu = int(node.status.allocatable.get("nvidia.com/gpu", "0")) # 已分配资源 allocated_cpu = node_pod_resources.get(node_name, {}).get("cpu", 0) / 1000.0 # 转换为 cores allocated_memory = node_pod_resources.get(node_name, {}).get("memory", 0) / (1024 ** 3) # 转换为 Gi allocated_gpu = node_pod_resources.get(node_name, {}).get("gpu", 0) # 可用资源 available_cpu = total_cpu - allocated_cpu available_memory = total_memory - allocated_memory available_gpu = total_gpu - allocated_gpu node_info = { # "node_name": node_name, # "node_status": status, # "node_role": roles_str, # "node_age": node_age, # "node_version": k8s_version, # "node_internalip": internal_ip, # "node_externalip": external_ip, # "node_osversion": os_image, # "node_kernelversion": kernel_version, # "node_containeruntime": container_runtime, # "node_labels": kyy_labels, # "node_cpurate": cpu_usage, # "node_memrate": memory_usage, # 新增资源信息 "node_total_cpu": f"{total_cpu:.2f}核", "node_allocated_cpu": f"{allocated_cpu:.2f}核", "node_available_cpu": f"{available_cpu:.2f}核", "node_cpu_usage_percent": f"{(allocated_cpu / total_cpu * 100):.1f}%" if total_cpu > 0 else "0%", "node_total_memory": f"{total_memory:.2f}Gi", "node_allocated_memory": f"{allocated_memory:.2f}Gi", "node_available_memory": f"{available_memory:.2f}Gi", "node_memory_usage_percent": f"{(allocated_memory / total_memory * 100):.1f}%" if total_memory > 0 else "0%", "node_total_gpu": total_gpu, "node_allocated_gpu": allocated_gpu, "node_available_gpu": available_gpu, "node_gpu_usage_percent": f"{(allocated_gpu / total_gpu * 100):.1f}%" if total_gpu > 0 else "0%" } rows.append(node_info) result = { "total": len(rows), "rows": rows } print(f"=== node_info={result}") return result except Exception as e: import traceback traceback.print_exc() print(f"获取节点信息失败: {e}") raise e # 辅助函数：解析资源值 def parse_resource_value(value: str) -> float: """解析 Kubernetes 资源值（如 "1.5", "500m", "2Gi"）为统一单位""" if not value: return 0.0 # 处理 CPU (cores 或 millicores) if value.endswith('m'): return float(value[:-1]) / 1000.0 # 转换为 cores elif re.match(r'^\d+(\.\d+)?$', value): return float(value) # 已经是 cores # 处理内存 (Ki, Mi, Gi, Ti) elif value.endswith('Ki'): return float(value[:-2]) / (1024 ** 1) # 转换为 Gi elif value.endswith('Mi'): return float(value[:-2]) / (1024 ** 2) elif value.endswith('Gi'): return float(value[:-2]) elif value.endswith('Ti'): return float(value[:-2]) * 1024 return float(value) # 默认按原单位返回 def get_pod_info(kubeconfig): try: # config.load_kube_config() kubeconfig = yaml.safe_load(kubeconfig) config.load_kube_config_from_dict(kubeconfig) v1 = client.CoreV1Api() api_client = client.ApiClient() namespaces = v1.list_namespace().items non_system_namespaces = [ns.metadata.name for ns in namespaces if not ns.metadata.name.startswith(('kube-', 'default', 'local', 'ingress-'))] rows = [] for namespace in non_system_namespaces: pods = v1.list_namespaced_pod(namespace).items pod_metrics_path = f"/apis/metrics.k8s.io/v1beta1/namespaces/{namespace}/pods" pod_metrics_response = api_client.call_api( pod_metrics_path, 'GET', auth_settings=['BearerToken'], response_type='object')[0] pod_metrics = {pod['metadata']['name']: pod.get("containers",[{}])[0].get('usage', {}) for pod in pod_metrics_response.get('items', [])} # debug(f"### pod_metrics_response={pod_metrics_response}") for pod in pods: pod_name = pod.metadata.name if pod.status.container_statuses: ready_count = sum(1 for cs in pod.status.container_statuses if cs.ready) else: ready_count = 0 ready_status = f"{ready_count}/{len(pod.spec.containers)}" readiness_conditions = [{"type": cond.type, "status": cond.status} for cond in pod.status.conditions if cond.type == "Ready"] phase = pod.status.phase restart_count = sum(cs.restart_count for cs in pod.status.container_statuses) if pod.status.container_statuses else 0 running_time = time.time() - pod.metadata.creation_timestamp.timestamp() pod_age = running_time pod_ip = pod.status.pod_ip if pod.status.pod_ip else "Unknown" node_name = pod.spec.node_name if pod.spec.node_name else "Pod未被调度到节点" nominated_node = pod.status.nominated_node_name if pod.status.nominated_node_name else "无" if phase == "Pending": pod_ip = "Pending状态,未分配 IP" node_name = "Pending状态,未分配节点" nominated_node = "Pending状态,未分配节点" # ✅ 提取容器的资源限制（limits） cpu_limit = "未设置" memory_limit = "未设置" gpu_limit = "未设置" if pod.spec.containers: container = pod.spec.containers[0] # 假设只取第一个容器 if container.resources and container.resources.limits: limits = container.resources.limits cpu_limit = limits.get("cpu", "未设置") memory_limit = limits.get("memory", "未设置") gpu_limit = limits.get("nvidia.com/gpu", "未设置") # 只支持 NVIDIA GPU # ✅ 获取 metrics 数据（已有逻辑不变） cpu_usage = pod_metrics.get(pod_name, {}).get('cpu', 'undefined') if cpu_usage and isinstance(cpu_usage, str): cpu_usage = int(cpu_usage.replace("n", "")) cpu_usage = f'{(cpu_usage / 1000000 / 1000):.3f}核' memory_usage = pod_metrics.get(pod_name, {}).get('memory', 'undefined') if memory_usage and isinstance(memory_usage, str): memory_usage = int(memory_usage.replace("Ki", "")) memory_usage = f"{(memory_usage / 1024):.3f}Mi" if phase in ["Pending", "Succeeded", "Failed"]: cpu_usage = "Pod未运行,无资源使用数据" memory_usage = "Pod未运行,无资源使用数据" # ✅ 新增 GPU 使用情况字段（暂时用占位符） gpu_usage = "0%" # 如果你有 DCGM / Prometheus 可替换为实际值 pod_info = { "pod_namespace": namespace, "pod_name": pod_name, "pod_ready": ready_status, "pod_running": phase, "pod_restart": str(restart_count), "pod_age": pod_age, "pod_ip": pod_ip, "pod_node": node_name, "pod_nominated_node": nominated_node, "pod_cpurate": cpu_usage, "pod_memrate": memory_usage, # ✅ 新增字段 "pod_cpu_limit": cpu_limit, "pod_memory_limit": memory_limit, "pod_gpu_limit": gpu_limit, "pod_gpu_usage": gpu_usage, } rows.append(pod_info) result = { "total": len(rows), "rows": rows } return result except Exception as e: raise "获取Pod信息失败: %s" % e.reason if __name__ == "__main__": kubeconfig = """apiVersion: v1 clusters: - cluster: certificate-authority-data: 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 server: https://192.168.0.3:6443 name: kubernetes contexts: - context: cluster: kubernetes user: kubernetes-admin name: kubernetes-admin@kubernetes current-context: kubernetes-admin@kubernetes kind: Config preferences: {} users: - name: kubernetes-admin user: client-certificate-data: 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 client-key-data: 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""" # 加载配置 # kubeconfig = yaml.safe_load(kubeconfig) # config.load_kube_config_from_dict(kubeconfig) # 测试获取节点信息 # try: # node_info = get_node_info(kubeconfig) # print(json.dumps(node_info, indent=4, ensure_ascii=False)) # except Exception as e: # print(f"Error: {e}") try: pod_info = get_pod_info(kubeconfig) print(json.dumps(pod_info, indent=4, ensure_ascii=False)) except Exception as e: print(f"Error: {e}")