回收闲置孤儿云资源并优化预留与计费模型
典型可挽回区间 5%~30% · 所属类别 🧾 财务与税务合规失血
| Provider | Resource count | Monthly spend | Annual spend | Spend share pct | Monthly wasted spend | Waste rate pct | Recoverable cash |
|---|---|---|---|---|---|---|---|
| AWS | 16 | 140,580.00 | 1,686,960.00 | 40.30 | 59,960.00 | 42.65 | 59,960.00 |
| Tencent Cloud | 11 | 59,930.00 | 719,160.00 | 17.18 | 34,040.00 | 56.80 | 34,040.00 |
| Aliyun | 13 | 57,500.00 | 690,000.00 | 16.48 | 25,950.00 | 45.13 | 25,950.00 |
| GCP | 4 | 33,700.00 | 404,400.00 | 9.66 | 29,800.00 | 88.43 | 29,800.00 |
| Volcengine | 4 | 30,600.00 | 367,200.00 | 8.77 | 28,450.00 | 92.97 | 28,450.00 |
| Azure | 4 | 26,500.00 | 318,000.00 | 7.60 | 15,700.00 | 59.25 | 15,700.00 |
仅展示前 10 行，完整明细可通过开放 API 沙盒获取
| Environment | Resource count | Monthly spend | Spend share pct | Monthly wasted spend | Waste rate pct | Recoverable cash |
|---|---|---|---|---|---|---|
| Production | 28 | 217,160.00 | 62.26 | 75,090.00 | 34.58 | 75,090.00 |
| Testing | 8 | 62,300.00 | 17.86 | 55,670.00 | 89.36 | 55,670.00 |
| Development | 9 | 48,890.00 | 14.02 | 48,740.00 | 99.69 | 48,740.00 |
| Staging | 7 | 20,460.00 | 5.87 | 14,400.00 | 70.38 | 14,400.00 |
| Category | Resource count | Monthly spend | Spend share pct | Monthly wasted spend | Waste rate pct | Recoverable cash |
|---|---|---|---|---|---|---|
| ECS/EC2 Compute | 25 | 219,000.00 | 62.78 | 123,430.00 | 56.36 | 123,430.00 |
| OSS/S3 Storage | 6 | 53,400.00 | 15.31 | 34,710.00 | 65.00 | 34,710.00 |
| RDS Database | 5 | 37,600.00 | 10.78 | 7,550.00 | 20.08 | 7,550.00 |
| NAT/EIP Bandwidth | 5 | 23,640.00 | 6.78 | 13,540.00 | 57.28 | 13,540.00 |
| EBS/CBS Disk | 8 | 10,120.00 | 2.90 | 10,120.00 | 100.00 | 10,120.00 |
| SLB/ALB Gateway | 3 | 5,050.00 | 1.45 | 4,550.00 | 90.10 | 4,550.00 |
| Department | Resource count | Monthly spend | Spend share pct | Monthly wasted spend | Waste rate pct | Recoverable cash |
|---|---|---|---|---|---|---|
| 算法研发部 | 8 | 96,400.00 | 27.64 | 83,220.00 | 86.33 | 83,220.00 |
| 数据平台部 | 7 | 62,450.00 | 17.90 | 26,550.00 | 42.51 | 26,550.00 |
| 运维基础设施部 | 11 | 61,690.00 | 17.69 | 38,140.00 | 61.83 | 38,140.00 |
| 商业化广告部 | 8 | 60,720.00 | 17.41 | 21,570.00 | 35.52 | 21,570.00 |
| 核心架构部 | 10 | 36,130.00 | 10.36 | 7,960.00 | 22.03 | 7,960.00 |
| 移动端业务部 | 8 | 31,420.00 | 9.01 | 16,460.00 | 52.39 | 16,460.00 |
| Leakage type | Label | Amount | Annual amount | Share pct | Affected count | Desc | Evidence |
|---|---|---|---|---|---|---|---|
| orphaned_disks_idle_compute | 🧟 僵尸孤儿资源空转扣费 (Orphaned Disks & Idle Compute) | 98,060.00 | 1,176,720.00 | 50.57 | 17 | 未挂载绑定的独立云硬盘 (Unattached EBS/CBS) 持续按容量秒级扣费，离职人员废弃测试机与空载实例 C… | AWS/阿里云未挂载云硬盘与孤儿弹性公网 IP (EIP) 持续计费，测试环境多台 16核机器整月 CPU 均值低于 … |
| overprovisioned_instance_sizing | 🐘 规格严重过度超配虚胖 (Over-provisioned Instance Sizing) | 46,480.00 | 557,760.00 | 23.97 | 11 | 架构师为应对偶发毛刺盲目按峰值 3~5 倍选用高配规格机型，常态 CPU<15% 且内存<20%，形成长期固定算力沉没… | 核心业务与中间件实例多选用 32核/64G 或 64核/128G 机型，日常算力水线极低，算力产能严重过剩。 |
| unarchived_cold_storage | 🧊 高价冷数据未分级归档 (Unarchived Cold Storage) | 34,710.00 | 416,520.00 | 17.90 | 6 | 海量日志备份、历史归档附件与历史分析冷数据未配置生命周期转储规则，长期占用昂贵高频标准型对象存储 (Standard … | 对象存储桶中超 90 天从未被检索读取的历史数据占比超 60%，仍持续支付标准型高昂存储单价。 |
| unhedged_on_demand_workloads | 💸 基础负载按需计费裸奔无预留 (Unhedged On-Demand Workloads) | 14,650.00 | 175,800.00 | 7.56 | 6 | 7×24 小时长期稳态生产底座与主从数据库完全按需 (On-Demand) 裸奔计费，未采购 1~3 年期 Savin… | 核心稳态集群按需计费支出占比过高，相比预留承诺折扣每月白白多支付 40%~72% 的溢价差额。 |
仅展示前 10 行，完整明细可通过开放 API 沙盒获取
| Resource id | Resource name | Cloud provider | Resource type | Environment | Department | Monthly spend | Cpu utilization |
|---|---|---|---|---|---|---|---|
| i-aws-ec2-001 | LLM-FineTune-Trainer-01 | AWS | ECS/EC2 Compute | Testing | 算法研发部 | 28,500.00 | 1.80 |
| i-ali-ecs-002 | Core-Order-Service-Prod-01 | Aliyun | ECS/EC2 Compute | Production | 核心架构部 | 3,200.00 | 68.50 |
| d-aws-ebs-003 | DataLake-Staging-Backup-Disk | AWS | EBS/CBS Disk | Staging | 数据平台部 | 1,850.00 | 0.00 |
| i-tc-cvm-004 | Recommendation-Inference-Node-A | Tencent Cloud | ECS/EC2 Compute | Production | 算法研发部 | 14,500.00 | 11.20 |
| oss-ali-005 | User-Raw-Logs-Bucket-2024 | Aliyun | OSS/S3 Storage | Production | 运维基础设施部 | 9,600.00 | 0.00 |
| rds-aws-006 | Payment-Gateway-Master-DB | AWS | RDS Database | Production | 核心架构部 | 12,800.00 | 58.20 |
| nat-tc-007 | Cross-AZ-NAT-Gateway-VPC01 | Tencent Cloud | NAT/EIP Bandwidth | Production | 运维基础设施部 | 6,800.00 | 0.00 |
| i-gcp-gce-008 | Crawler-Cluster-Worker-Dev-03 | GCP | ECS/EC2 Compute | Development | 商业化广告部 | 4,200.00 | 2.40 |
| d-tc-cbs-009 | Old-Jenkins-Artifact-Vol | Tencent Cloud | EBS/CBS Disk | Testing | 核心架构部 | 1,350.00 | 0.00 |
| i-az-vm-010 | BI-Reporting-ETL-Node-02 | Azure | ECS/EC2 Compute | Production | 数据平台部 | 16,200.00 | 12.00 |
仅展示前 10 行，完整明细可通过开放 API 沙盒获取
| Core high utilization count | 5 |
| Agile efficient micro count | 9 |
| Bloated overprovisioned count | 18 |
| Zombie idle bleeder count | 20 |
| Resource id | Resource name | Cloud provider | Resource type | Environment | Department | Instance type | Monthly spend |
|---|---|---|---|---|---|---|---|
| i-aws-ec2-001 | LLM-FineTune-Trainer-01 | AWS | ECS/EC2 Compute | Testing | 算法研发部 | g5.12xlarge | 28,500.00 |
| i-gcp-gce-042 | LLM-Prompt-Engineering-DevBox | GCP | ECS/EC2 Compute | Development | 算法研发部 | a2-highgpu-1g | 19,500.00 |
| i-vol-ecs-023 | Feature-Store-Offline-Trainer | Volcengine | ECS/EC2 Compute | Development | 算法研发部 | ecs.c3i.16xlarge | 16,800.00 |
| s3-aws-013 | Historical-Audit-Logs-Archive | AWS | OSS/S3 Storage | Production | 运维基础设施部 | S3-Standard-120TB | 14,400.00 |
| i-vol-ecs-012 | Ad-Model-Validation-Node-05 | Volcengine | ECS/EC2 Compute | Testing | 商业化广告部 | ecs.g3i.8xlarge | 8,600.00 |
| i-az-vm-010 | BI-Reporting-ETL-Node-02 | Azure | ECS/EC2 Compute | Production | 数据平台部 | Standard_E32s_v5 | 16,200.00 |
| i-tc-cvm-004 | Recommendation-Inference-Node-A | Tencent Cloud | ECS/EC2 Compute | Production | 算法研发部 | GN7.8XLARGE | 14,500.00 |
| s3-aws-039 | Model-Checkpoints-DeepCold | AWS | OSS/S3 Storage | Testing | 算法研发部 | S3-Standard-90TB | 10,800.00 |
| oss-ali-005 | User-Raw-Logs-Bucket-2024 | Aliyun | OSS/S3 Storage | Production | 运维基础设施部 | Standard-Storage-80TB | 9,600.00 |
| i-tc-cvm-038 | BigData-Presto-Coordinator | Tencent Cloud | ECS/EC2 Compute | Production | 数据平台部 | M5.8XLARGE | 9,800.00 |
仅展示前 10 行，完整明细可通过开放 API 沙盒获取
| Resource id | Resource name | Cloud provider | Resource type | Environment | Department | Instance type | Monthly spend |
|---|---|---|---|---|---|---|---|
| i-aws-ec2-001 | LLM-FineTune-Trainer-01 | AWS | ECS/EC2 Compute | Testing | 算法研发部 | g5.12xlarge | 28,500.00 |
| i-ali-ecs-002 | Core-Order-Service-Prod-01 | Aliyun | ECS/EC2 Compute | Production | 核心架构部 | ecs.g7.4xlarge | 3,200.00 |
| d-aws-ebs-003 | DataLake-Staging-Backup-Disk | AWS | EBS/CBS Disk | Staging | 数据平台部 | gp3-2000GB | 1,850.00 |
| i-tc-cvm-004 | Recommendation-Inference-Node-A | Tencent Cloud | ECS/EC2 Compute | Production | 算法研发部 | GN7.8XLARGE | 14,500.00 |
| oss-ali-005 | User-Raw-Logs-Bucket-2024 | Aliyun | OSS/S3 Storage | Production | 运维基础设施部 | Standard-Storage-80TB | 9,600.00 |
| rds-aws-006 | Payment-Gateway-Master-DB | AWS | RDS Database | Production | 核心架构部 | db.r6g.4xlarge | 12,800.00 |
| nat-tc-007 | Cross-AZ-NAT-Gateway-VPC01 | Tencent Cloud | NAT/EIP Bandwidth | Production | 运维基础设施部 | NAT-Ultra-Large | 6,800.00 |
| i-gcp-gce-008 | Crawler-Cluster-Worker-Dev-03 | GCP | ECS/EC2 Compute | Development | 商业化广告部 | n2-standard-16 | 4,200.00 |
| d-tc-cbs-009 | Old-Jenkins-Artifact-Vol | Tencent Cloud | EBS/CBS Disk | Testing | 核心架构部 | CBS-SSD-1500GB | 1,350.00 |
| i-az-vm-010 | BI-Reporting-ETL-Node-02 | Azure | ECS/EC2 Compute | Production | 数据平台部 | Standard_E32s_v5 | 16,200.00 |
仅展示前 10 行，完整明细可通过开放 API 沙盒获取
| Executive summary | 本次全盘云基础设施审计共穿透覆盖企业在用的 52 项核心云资产，全盘月度云支出规模达 ¥348,810.00 元 (折算年度支出 ¥4,185,720.00 元)。审计穿透发现全盘综合闲置浪费率高达 55.59%，月度纯闲置与配置浪费总额… |
| Health commentary | 企业整体云上支出处于【🚨 算力失血黑洞 (Severe Cloud Bleeder)】阶段。主要症结在于早期为追求交付速度普遍采取‘冗余超配买心安’策略，开发测试环境与离职员工创建的虚拟机缺乏自动化全生命周期销毁机制；同时大量 7×24 … |
| Immediate actions | ["P0 紧急清退：7 天内强制批量销毁全部 10 块孤儿云硬盘与废弃测试机，直接阻断 ¥98,060.00/月 纯失血；", "P1 架构降配：30 天内对 11 台超配算力执行弹性减半降档并开启冷数据生命周期转储，月省 ¥81,190… |
| Projected impact | 云上降本节省的每一分钱 100% 毫无损耗直转为企业的纯现金营业利润 (EBITDA)。全盘落地 5 大动态 What-If 推演优化后，预计年度可直接释放纯现金 EBITDA 营业纯利 ¥1,977,780.00 元，企业年度云总账单将… |
| Playbook strategy | 全面贯彻‘P0 即刻止血清孤儿 → P1 架构降配冷归档 → P2 稳态承诺统签锁底价’三阶麦肯锡战役模型，固化跨部门云成本分摊与考核机制。 |
| Audit confidence score | 98.50 |
52
cloud_finops
✅ 公共沙盒端点已开放无需注册、无需 API Key，把下面这段 curl 直接粘到终端，就能拿到与本页结构完全一致的 JSON（返回体包含 engine_name / record_count 与全部分析区块）。
curl -X POST 'https://data-copilot.whfc.cc/api/v1/sandbox/cloud_finops' \
-H 'Content-Type: application/json' \
-d '{"use_sample": true}'同一失血大类下，其它引擎的真实样例输出（免注册查看）：
页面来源：https://data-copilot.whfc.cc/demo/cloud_finops
Data Copilot · AI 跨境经营失血审计 —— 把财务报表里看不见的流失，还原成可以逐笔挽回的现金。
商务对接与报告转发：bd@data-copilot.whfc.cc · 全部 54 个审计引擎 · 开放 API 开发者门户