• Cost Management Reader 角色
• Monitoring Reader 角色
• 订阅/资源组上的 Reader 角色

验证命令：

az --version
az account show
az extension show --name costmanagement
azqr version

步骤 1：加载最佳实践

获取 Azure 成本优化最佳实践，为建议提供依据：

// Use Azure MCP best practices tool
mcp_azure_mcp_get_azure_bestpractices({
  intent: "Get cost optimization best practices",
  command: "get_bestpractices",
  parameters: { resource: "cost-optimization", action: "all" }
})

步骤 1.5：Redis 特定分析（条件性）

如果用户特别请求 Redis 成本优化，请使用专门的 Redis 技能：

📋 参考：Azure Redis 成本优化

何时使用 Redis 特定分析：

• 用户提到 "Redis"、"Azure Cache for Redis" 或 "Azure Managed Redis"
• 重点是 Redis 资源优化，而非通用订阅分析
• 用户需要 Redis 特定建议（SKU 降级、故障缓存等）

关键能力：

• 交互式订阅筛选（前缀、ID 或 "所有订阅"）
• Redis 特定优化规则（故障缓存、过大规模层级、缺失标签）
• 用于 Redis 成本分析的预构建报告模板
• 使用 redis_list 命令

可用的报告模板：

• 订阅级别 Redis 摘要
• 详细的 Redis 缓存分析

注意：对于通用的订阅范围成本优化（包括 Redis），请继续执行步骤 2。对于仅关注 Redis 的分析，请遵循 Redis 特定参考文档中的说明。

步骤 1.6：选择分析范围（针对 Redis 特定分析）

如果执行 Redis 成本优化，请让用户选择分析范围：

向用户提示以下选项：

1. 特定订阅 ID - 分析单个订阅
2. 订阅名称 - 使用显示名称而非 ID
3. 订阅前缀 - 分析以某个前缀开头的所有订阅（例如，"CacheTeam"）
4. 我的所有订阅 - 扫描所有可访问的订阅
5. 租户范围 - 分析整个组织

在继续执行步骤 2 之前，等待用户响应。

步骤 2：运行 Azure Quick Review

运行 azqr 以查找孤立资源（即时节省成本）：

📋 参考：Azure Quick Review - 运行 azqr 扫描的详细说明

// Use Azure MCP extension_azqr tool
extension_azqr({
  subscription: "<SUBSCRIPTION_ID>",
  "resource-group": "<RESOURCE_GROUP>"  // optional
})

在 azqr 结果中需要关注的内容：

• 孤立资源：未挂载的磁盘、未使用的网卡、空闲的 NAT 网关
• 过度配置的资源：过长的保留期、过大的 SKU
• 缺失的成本标签：没有正确成本分配的资源

注意：Azure Quick Review 参考文档包含创建筛选配置、将输出保存到 output/ 文件夹以及解释成本优化结果的说明。

步骤 3：发现资源

为了高效地进行跨订阅资源发现，请使用 Azure Resource Graph。有关孤立资源检测和成本优化模式，请参阅 Azure Resource Graph 查询。

使用 Azure MCP 工具或 CLI 列出订阅中的所有资源：

# Get subscription info
az account show

# List all resources
az resource list --subscription "<SUBSCRIPTION_ID>" --resource-group "<RESOURCE_GROUP>"

# Use MCP tools for specific services (preferred):
# - Storage accounts, Cosmos DB, Key Vaults: use Azure MCP tools
# - Redis caches: use mcp_azure_mcp_redis tool (see ./references/azure-redis.md)
# - Web apps, VMs, SQL: use az CLI commands

步骤 4：查询实际成本

从 Azure Cost Management API 获取实际成本数据（最近 30 天）：

创建成本查询文件：

创建 temp/cost-query.json，内容如下：

{
  "type": "ActualCost",
  "timeframe": "Custom",
  "timePeriod": {
    "from": "<START_DATE>",  
    "to": "<END_DATE>"
  },
  "dataset": {
    "granularity": "None",
    "aggregation": {
      "totalCost": {
        "name": "Cost",
        "function": "Sum"
      }
    },
    "grouping": [
      {
        "type": "Dimension",
        "name": "ResourceId"
      }
    ]
  }
}

必需操作：计算 <START_DATE>（30 天前）和 <END_DATE>（今天）的 ISO 8601 格式（例如 2025-11-03T00:00:00Z）。

执行成本查询：

# Create temp folder
New-Item -ItemType Directory -Path "temp" -Force

# Query using REST API (more reliable than az costmanagement query)
az rest --method post `
  --url "https://management.azure.com/subscriptions/<SUBSCRIPTION_ID>/resourceGroups/<RESOURCE_GROUP>/providers/Microsoft.CostManagement/query?api-version=2023-11-01" `
  --body '@temp/cost-query.json'

重要提示： 将查询结果保存到 output/cost-query-result<timestamp>.json 以用于审计跟踪。

步骤 5：验证定价

使用 fetch_webpage 从官方 Azure 定价页面获取当前定价：

// Validate pricing for key services
fetch_webpage({
  urls: ["https://azure.microsoft.com/en-us/pricing/details/container-apps/"],
  query: "pricing tiers and costs"
})

需要验证定价的关键服务：

• Container Apps: https://azure.microsoft.com/pricing/details/container-apps/
• Virtual Machines: https://azure.microsoft.com/pricing/details/virtual-machines/
• App Service: https://azure.microsoft.com/pricing/details/app-service/
• Log Analytics: https://azure.microsoft.com/pricing/details/monitor/

重要提示： 检查免费层额度 - 许多 Azure 服务都有慷慨的免费额度，这可能解释了 $0 成本。

步骤 6：收集使用率指标

查询 Azure Monitor 获取使用率数据（最近 14 天），以支持调整规模的建议：

# Calculate dates for last 14 days
$startTime = (Get-Date).AddDays(-14).ToString("yyyy-MM-ddTHH:mm:ssZ")
$endTime = Get-Date -Format "yyyy-MM-ddTHH:mm:ssZ"

# VM CPU utilization
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "Percentage CPU" `
  --interval PT1H `
  --aggregation Average `
  --start-time $startTime `
  --end-time $endTime

# App Service Plan utilization
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "CpuTime,Requests" `
  --interval PT1H `
  --aggregation Total `
  --start-time $startTime `
  --end-time $endTime

# Storage capacity
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "UsedCapacity,BlobCount" `
  --interval PT1H `
  --aggregation Average `
  --start-time $startTime `
  --end-time $endTime

步骤 7：生成优化报告

在 output/ 文件夹中创建全面的成本优化报告：

使用 create_file 工具，路径为 output/costoptimizereport<YYYYMMDD_HHMMSS>.md：

报告结构：

# Azure 成本优化报告
**生成时间**：<timestamp>

## 执行摘要
- 月度总成本：$X (💰 实际数据)
- 主要成本驱动因素：[列出前 3 个资源，附带 Azure 门户链接]

## 成本细分
[按成本排序的前 10 个资源表格，包含 Azure 门户链接]

## 免费层分析
[在免费层额度内运行、显示 $0 成本的资源]

## 孤立资源（即时节省）
[来自 azqr - 可以立即删除的资源]
- 资源名称及门户链接 - 每月节省 $X

## 优化建议

### 优先级 1：高影响，低风险
[示例：删除孤立资源]
- 💰 实际成本：每月 $X
- 📊 预计节省：每月 $Y
- 要执行的命令（附警告）

### 优先级 2：中等影响，中等风险
[示例：将 VM 从 D4s_v5 调整为 D2s_v5]
- 💰 实际基准：D4s_v5，每月 $X
- 📈 实际指标：CPU 8%，内存 30%
- 💵 已验证定价：D4s_v5 $Y/小时，D2s_v5 $Z/小时
- 📊 预计节省：每月 $S
- 要执行的命令

### 优先级 3：长期优化
[示例：预留实例、存储分层]

## 预计总节省
- 月度：$X
- 年度：$Y

## 实施命令
[附有批准警告的安全命令]

## 验证附录

### 数据源和文件
- **成本查询结果**：`output/cost-query-result<timestamp>.json`
  - 来自 Azure Cost Management API 的原始成本数据
  - 证明报告生成时实际成本的审计跟踪
  - 至少保留 12 个月以进行历史比较
  - 包含分析期间每个资源的确切成本
- **定价来源**：[Azure 定价页面链接]
- **免费层额度**：[适用的额度]

> **注意**：`temp/cost-query.json` 文件（如果存在）是临时查询模板，可以安全删除。所有永久性审计数据都在 `output/` 文件夹中。

门户链接格式：

https://portal.azure.com/#@<TENANT_ID>/resource/subscriptions/<SUBSCRIPTION_ID>/resourceGroups/<RESOURCE_GROUP>/providers/<RESOURCE_PROVIDER>/<RESOURCE_TYPE>/<RESOURCE_NAME>/overview

步骤 8：保存审计跟踪

保存所有成本查询结果以供验证：

使用 create_file 工具，路径为 output/cost-query-result<YYYYMMDD_HHMMSS>.json：

{
  "timestamp": "<ISO_8601>",
  "subscription": "<SUBSCRIPTION_ID>",
  "resourceGroup": "<RESOURCE_GROUP>",
  "queries": [
    {
      "queryType": "ActualCost",
      "timeframe": "MonthToDate",
      "query": { },
      "response": { }
    }
  ]
}

步骤 9：清理临时文件

报告生成后，删除临时查询文件和文件夹：

# Delete entire temp folder (no longer needed)
Remove-Item -Path "temp" -Recurse -Force -ErrorAction SilentlyContinue

注意：temp/cost-query.json 文件仅在 API 执行期间需要。实际的查询和结果保存在 output/cost-query-result*.json 中以用于审计目的。

输出

该技能生成：

1. 成本优化报告 (output/costoptimizereport<timestamp>.md)
   • 包含总成本和主要驱动因素的执行摘要
   • 带有 Azure 门户链接的详细成本细分
   • 基于实际数据和已验证定价的优先级建议及预计节省
   • 带有安全警告的实施命令
2. 成本查询结果 (output/cost-query-result<timestamp>.json)
   • 所有成本查询和响应的审计跟踪
   • 建议的验证证据

重要说明

数据分类

• 💰 实际数据 = 从 Azure Cost Management API 检索
• 📈 实际指标 = 从 Azure Monitor 检索
• 💵 已验证定价 = 从官方 Azure 定价页面检索
• 📊 预计节省 = 基于实际数据和已验证定价计算

最佳实践

• 始终首先查询实际成本 - 切勿估计或假设
• 从官方来源验证定价 - 考虑免费层额度
• 使用 REST API 进行成本查询（比 az costmanagement query 更可靠）
• 保存审计跟踪 - 包含所有查询和响应
• 为所有资源包含 Azure 门户链接
• 创建报告文件时使用 UTF-8 编码
• 对于成本 < $10/月的资源，强调操作改进而非财务节省
• 未经明确批准，切勿执行破坏性操作

常见陷阱

• 假设成本：始终从 Cost Management API 查询实际数据
• 忽略免费层：许多服务有慷慨的额度（例如，Container Apps：每月 180K vCPU 秒免费）
• 使用错误的日期范围：成本用 30 天，使用率用 14 天
• 门户链接错误：验证租户 ID 和资源 ID 格式
• 成本查询失败：使用带有 JSON 主体的 az rest，而不是 az costmanagement query

安全要求

• 删除资源前获得批准
• 先在非生产环境测试变更
• 提供用于验证的试运行命令
• 包含回滚程序
• 实施后监控影响

SDK 快速参考

• Redis 管理：.NET

每周安装量

102.6K

代码仓库

microsoft/githu…or-azure

GitHub 星标数

143

首次出现

2026年2月4日

安全审计

Gen Agent Trust HubPassSocketPassSnykPass

安装于

github-copilot102.5K

codex390

gemini-cli380

opencode363

kimi-cli351

amp351

Required Permissions:

• Cost Management Reader role
• Monitoring Reader role
• Reader role on subscription/resource group

Verification commands:

az --version
az account show
az extension show --name costmanagement
azqr version

Step 1: Load Best Practices

Get Azure cost optimization best practices to inform recommendations:

// Use Azure MCP best practices tool
mcp_azure_mcp_get_azure_bestpractices({
  intent: "Get cost optimization best practices",
  command: "get_bestpractices",
  parameters: { resource: "cost-optimization", action: "all" }
})

Step 1.5: Redis-Specific Analysis (Conditional)

If the user specifically requests Redis cost optimization , use the specialized Redis skill:

📋 Reference : Azure Redis Cost Optimization

When to use Redis-specific analysis:

• User mentions "Redis", "Azure Cache for Redis", or "Azure Managed Redis"
• Focus is on Redis resource optimization, not general subscription analysis
• User wants Redis-specific recommendations (SKU downgrade, failed caches, etc.)

Key capabilities:

• Interactive subscription filtering (prefix, ID, or "all subscriptions")
• Redis-specific optimization rules (failed caches, oversized tiers, missing tags)
• Pre-built report templates for Redis cost analysis
• Uses redis_list command

Report templates available:

• Subscription-level Redis summary
• Detailed Redis cache analysis

Note : For general subscription-wide cost optimization (including Redis), continue with Step 2. For Redis-only focused analysis, follow the instructions in the Redis-specific reference document.

Step 1.6: Choose Analysis Scope (for Redis-specific analysis)

If performing Redis cost optimization , ask the user to select their analysis scope:

Prompt the user with these options:

1. Specific Subscription ID - Analyze a single subscription
2. Subscription Name - Use display name instead of ID
3. Subscription Prefix - Analyze all subscriptions starting with a prefix (e.g., "CacheTeam")
4. All My Subscriptions - Scan all accessible subscriptions
5. Tenant-wide - Analyze entire organization

Wait for user response before proceeding to Step 2.

Step 2: Run Azure Quick Review

Run azqr to find orphaned resources (immediate cost savings):

📋 Reference : Azure Quick Review - Detailed instructions for running azqr scans

// Use Azure MCP extension_azqr tool
extension_azqr({
  subscription: "<SUBSCRIPTION_ID>",
  "resource-group": "<RESOURCE_GROUP>"  // optional
})

What to look for in azqr results:

• Orphaned resources: unattached disks, unused NICs, idle NAT gateways
• Over-provisioned resources: excessive retention periods, oversized SKUs
• Missing cost tags: resources without proper cost allocation

Note : The Azure Quick Review reference document includes instructions for creating filter configurations, saving output to the output/ folder, and interpreting results for cost optimization.

Step 3: Discover Resources

For efficient cross-subscription resource discovery, use Azure Resource Graph. See Azure Resource Graph Queries for orphaned resource detection and cost optimization patterns.

List all resources in the subscription using Azure MCP tools or CLI:

# Get subscription info
az account show

# List all resources
az resource list --subscription "<SUBSCRIPTION_ID>" --resource-group "<RESOURCE_GROUP>"

# Use MCP tools for specific services (preferred):
# - Storage accounts, Cosmos DB, Key Vaults: use Azure MCP tools
# - Redis caches: use mcp_azure_mcp_redis tool (see ./references/azure-redis.md)
# - Web apps, VMs, SQL: use az CLI commands

Step 4: Query Actual Costs

Get actual cost data from Azure Cost Management API (last 30 days):

Create cost query file:

Create temp/cost-query.json with:

{
  "type": "ActualCost",
  "timeframe": "Custom",
  "timePeriod": {
    "from": "<START_DATE>",  
    "to": "<END_DATE>"
  },
  "dataset": {
    "granularity": "None",
    "aggregation": {
      "totalCost": {
        "name": "Cost",
        "function": "Sum"
      }
    },
    "grouping": [
      {
        "type": "Dimension",
        "name": "ResourceId"
      }
    ]
  }
}

Action Required : Calculate <START_DATE> (30 days ago) and <END_DATE> (today) in ISO 8601 format (e.g., 2025-11-03T00:00:00Z).

Execute cost query:

# Create temp folder
New-Item -ItemType Directory -Path "temp" -Force

# Query using REST API (more reliable than az costmanagement query)
az rest --method post `
  --url "https://management.azure.com/subscriptions/<SUBSCRIPTION_ID>/resourceGroups/<RESOURCE_GROUP>/providers/Microsoft.CostManagement/query?api-version=2023-11-01" `
  --body '@temp/cost-query.json'

Important: Save the query results to output/cost-query-result<timestamp>.json for audit trail.

Step 5: Validate Pricing

Fetch current pricing from official Azure pricing pages using fetch_webpage:

// Validate pricing for key services
fetch_webpage({
  urls: ["https://azure.microsoft.com/en-us/pricing/details/container-apps/"],
  query: "pricing tiers and costs"
})

Key services to validate:

• Container Apps: https://azure.microsoft.com/pricing/details/container-apps/
• Virtual Machines: https://azure.microsoft.com/pricing/details/virtual-machines/
• App Service: https://azure.microsoft.com/pricing/details/app-service/
• Log Analytics: https://azure.microsoft.com/pricing/details/monitor/

Important : Check for free tier allowances - many Azure services have generous free limits that may explain $0 costs.

Step 6: Collect Utilization Metrics

Query Azure Monitor for utilization data (last 14 days) to support rightsizing recommendations:

# Calculate dates for last 14 days
$startTime = (Get-Date).AddDays(-14).ToString("yyyy-MM-ddTHH:mm:ssZ")
$endTime = Get-Date -Format "yyyy-MM-ddTHH:mm:ssZ"

# VM CPU utilization
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "Percentage CPU" `
  --interval PT1H `
  --aggregation Average `
  --start-time $startTime `
  --end-time $endTime

# App Service Plan utilization
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "CpuTime,Requests" `
  --interval PT1H `
  --aggregation Total `
  --start-time $startTime `
  --end-time $endTime

# Storage capacity
az monitor metrics list `
  --resource "<RESOURCE_ID>" `
  --metric "UsedCapacity,BlobCount" `
  --interval PT1H `
  --aggregation Average `
  --start-time $startTime `
  --end-time $endTime

Step 7: Generate Optimization Report

Create a comprehensive cost optimization report in the output/ folder:

Use thecreate_file tool with path output/costoptimizereport<YYYYMMDD_HHMMSS>.md:

Report Structure:

# Azure Cost Optimization Report
**Generated**: <timestamp>

## Executive Summary
- Total Monthly Cost: $X (💰 ACTUAL DATA)
- Top Cost Drivers: [List top 3 resources with Azure Portal links]

## Cost Breakdown
[Table with top 10 resources by cost, including Azure Portal links]

## Free Tier Analysis
[Resources operating within free tiers showing $0 cost]

## Orphaned Resources (Immediate Savings)
[From azqr - resources that can be deleted immediately]
- Resource name with Portal link - $X/month savings

## Optimization Recommendations

### Priority 1: High Impact, Low Risk
[Example: Delete orphaned resources]
- 💰 ACTUAL cost: $X/month
- 📊 ESTIMATED savings: $Y/month
- Commands to execute (with warnings)

### Priority 2: Medium Impact, Medium Risk
[Example: Rightsize VM from D4s_v5 to D2s_v5]
- 💰 ACTUAL baseline: D4s_v5, $X/month
- 📈 ACTUAL metrics: CPU 8%, Memory 30%
- 💵 VALIDATED pricing: D4s_v5 $Y/hr, D2s_v5 $Z/hr
- 📊 ESTIMATED savings: $S/month
- Commands to execute

### Priority 3: Long-term Optimization
[Example: Reserved Instances, Storage tiering]

## Total Estimated Savings
- Monthly: $X
- Annual: $Y

## Implementation Commands
[Safe commands with approval warnings]

## Validation Appendix

### Data Sources and Files
- **Cost Query Results**: `output/cost-query-result<timestamp>.json`
  - Raw cost data from Azure Cost Management API
  - Audit trail proving actual costs at report generation time
  - Keep for at least 12 months for historical comparison
  - Contains every resource's exact cost over the analysis period
- **Pricing Sources**: [Links to Azure pricing pages]
- **Free Tier Allowances**: [Applicable allowances]

> **Note**: The `temp/cost-query.json` file (if present) is a temporary query template and can be safely deleted. All permanent audit data is in the `output/` folder.

Portal Link Format:

https://portal.azure.com/#@<TENANT_ID>/resource/subscriptions/<SUBSCRIPTION_ID>/resourceGroups/<RESOURCE_GROUP>/providers/<RESOURCE_PROVIDER>/<RESOURCE_TYPE>/<RESOURCE_NAME>/overview

Step 8: Save Audit Trail

Save all cost query results for validation:

Use thecreate_file tool with path output/cost-query-result<YYYYMMDD_HHMMSS>.json:

{
  "timestamp": "<ISO_8601>",
  "subscription": "<SUBSCRIPTION_ID>",
  "resourceGroup": "<RESOURCE_GROUP>",
  "queries": [
    {
      "queryType": "ActualCost",
      "timeframe": "MonthToDate",
      "query": { },
      "response": { }
    }
  ]
}

Step 9: Clean Up Temporary Files

Remove temporary query files and folder after the report is generated:

# Delete entire temp folder (no longer needed)
Remove-Item -Path "temp" -Recurse -Force -ErrorAction SilentlyContinue

Note : The temp/cost-query.json file is only needed during API execution. The actual query and results are preserved in output/cost-query-result*.json for audit purposes.

Output

The skill generates:

1. Cost Optimization Report (output/costoptimizereport<timestamp>.md)

   • Executive summary with total costs and top drivers
   • Detailed cost breakdown with Azure Portal links
   • Prioritized recommendations with actual data and estimated savings
   • Implementation commands with safety warnings

2. Cost Query Results (output/cost-query-result<timestamp>.json)

   • Audit trail of all cost queries and responses
   • Validation evidence for recommendations

Important Notes

Data Classification

• 💰 ACTUAL DATA = Retrieved from Azure Cost Management API
• 📈 ACTUAL METRICS = Retrieved from Azure Monitor
• 💵 VALIDATED PRICING = Retrieved from official Azure pricing pages
• 📊 ESTIMATED SAVINGS = Calculated based on actual data and validated pricing

Best Practices

• Always query actual costs first - never estimate or assume
• Validate pricing from official sources - account for free tiers
• Use REST API for cost queries (more reliable than az costmanagement query)
• Save audit trail - include all queries and responses
• Include Azure Portal links for all resources
• Use UTF-8 encoding when creating report files
• For costs < $10/month, emphasize operational improvements over financial savings
• Never execute destructive operations without explicit approval

Common Pitfalls

• Assuming costs : Always query actual data from Cost Management API
• Ignoring free tiers : Many services have generous allowances (e.g., Container Apps: 180K vCPU-sec free/month)
• Using wrong date ranges : 30 days for costs, 14 days for utilization
• Broken Portal links : Verify tenant ID and resource ID format
• Cost query failures : Use az rest with JSON body, not az costmanagement query

Safety Requirements

• Get approval before deleting resources
• Test changes in non-production first
• Provide dry-run commands for validation
• Include rollback procedures
• Monitor impact after implementation

SDK Quick References

• Redis Management : .NET

Weekly Installs

102.6K

Repository

microsoft/githu…or-azure

GitHub Stars

143

First Seen

Feb 4, 2026

Security Audits

Gen Agent Trust HubPassSocketPassSnykPass

Installed on

github-copilot102.5K

codex390

gemini-cli380

opencode363

kimi-cli351

amp351