Datadog MCP Server 27315d08

MCP运维云端

提供方 dreamiurg · 包名 github

运维云端 stdio · 本地进程 无需鉴权 社区 / 第三方 Claude DesktopClaudeCursorWindsurf
查看接入方式

Datadog MCP Server 是一个提供117种只读工具的观测性平台,覆盖日志、APM、指标、RUM等所有观测支柱,专为AI助手设计,确保数据可视性同时防止误操作。

一句话选型结论

按资料字段整理

Datadog MCP Server 27315d08 属于运维云端类 MCP 服务器,stdio · 本地进程,无需鉴权,可直接接入 Claude Desktop、Claude、Cursor 等,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类运维云端 传输方式stdio · 本地进程 鉴权无需鉴权 部署方式本地部署
GitHub 星标 1 仓库热度,无公开仓库时显示未标注
工具数 112 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

运维云端
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量112暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

观测性平台TypeScript本地部署只读工具AI集成日志管理APM追踪实时用户监控

安装与配置

stdio · 本地进程 无需鉴权
客户端在本机拉起子进程,数据不出本机
npx github:dreamiurg/datadog-mcp --help

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 Node.js,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后客户端会列出约 112 个工具,随便调一个只读工具确认能返回结果。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
云资源查询

查实例、存储、账单等云上资源状态。

部署与发布

把部署、重启、回滚这类操作接进对话。

监控与告警

拉取监控指标,异常时让模型帮你判断。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 资料标注 112 个工具,覆盖面较广,能满足多种调用。
  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude Desktop、Claude、Cursor,换客户端不用重新研究配置。
  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在运维云端场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Datadog MCP Server 27315d08当前 112 stdio · 本地进程 本地部署 有 2026-05-17
lobehub 未标注 stdio · 本地进程 本地部署 有 2026-04-11
TrendRadar 未标注 stdio · 本地进程 本地部署 有 2026-04-13
Ui Tars Desktop 8a825770 未标注 stdio · 本地进程 本地部署 有 2026-05-17
ruflo 未标注 stdio · 本地进程 本地部署 有 2026-04-13
FastGPT 未标注 stdio · 本地进程 本地部署 有 2026-04-11

服务器档案

运维云端
字段来自项目仓库与公开索引,中文说明由本站标注
提供方dreamiurg上游标注的维护方 dreamiurg
分类运维云端本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时github · Node.js本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

27.6 KB
来自项目仓库与上游目录的原文,未做删改

Datadog MCP Server

> The Datadog MCP server your AI deserves. 117 read-only tools. Every observability pillar. Zero writes. Zero regrets.


Why This One?

Look, we get it. You searched "datadog mcp server" and got seven results. You're wondering which one to pick. Let us save you approximately forty-five seconds of reading other READMEs that say "supports monitors and dashboards" like that's supposed to impress someone.

Feature Us (117 tools) Project 1 Project 2 Project 3
Observability Core
Logs (search + aggregate + pipelines + indexes) 7 tools 1 tool 0-1 tool 0-1 tool
APM (traces, spans, services, dependencies) 7 tools 1 tool 0 tools 0 tools
Metrics (query, metadata, volumes, tags) 6 tools 1 tool 0-1 tool 0 tools
RUM (events, apps, aggregation) 3 tools 3 tools 0 tools 0 tools
Alerting & Reliability
Monitors (list, get, search) 4 tools 1 tool 0-1 tool 0-1 tool
Dashboards (get, list, collections) 3 tools 1 tool 0-1 tool 0-1 tool
SLOs (list, get, history, search, corrections) 5 tools 0 tools 0-1 tool 0-1 tool
Synthetics (tests, results, locations) 5 tools 0 tools 0-1 tool 0 tools
Events & Incidents 7 tools 2 tools 0-1 tool 0 tools
Downtimes 2 tools 1 tool 0-1 tool 0 tools
Infrastructure & Security
Infrastructure (hosts, tags, containers, processes) 8 tools 2 tools 0-1 tool 0-1 tool
Security (findings, posture, signals, rules, scanning) 10 tools 0 tools 0-1 tool 0 tools
Network Monitoring 2 tools 0 tools 0 tools 0 tools
Cloud Integrations (AWS/GCP/Azure/Cloudflare/Confluent) 5 tools 0 tools 0 tools 0 tools
Platform & DevOps
Service Catalog 3 tools 0 tools 0-1 tool 0 tools
CI Visibility (pipelines + tests) 4 tools 0 tools 0-1 tool 0 tools
Organization (teams, users, auth) 5 tools 0 tools 0-1 tool 0 tools
Error Tracking 3 tools 0 tools 0 tools 0 tools
Database Monitoring 2 tools 0 tools 0 tools 0 tools
Workflows & Automation 2 tools 0 tools 0 tools 0 tools
DORA Metrics 1 tool 0 tools 0 tools 0 tools
Governance & Cost
Audit Trail 2 tools 0 tools 0 tools 0 tools
Usage & Billing 4 tools 0 tools 0 tools 0 tools
Cloud Cost Management 1 tool 0 tools 0 tools 0 tools
Access Control (keys, mappings, policies) 4 tools 0 tools 0 tools 0 tools
Notebooks 3 tools 0 tools 0 tools 0 tools
Fleet Management 1 tool 0 tools 0 tools 0 tools
Webhooks 1 tool 0 tools 0 tools 0 tools
Monitor Notifications 1 tool 0 tools 0 tools 0 tools
IP Ranges & Config 1 tool 0 tools 0 tools 0 tools
Total 117 ~20 ~13 ~29
Accidentally deletes your monitors No No No No
Can mute your hosts at 3 AM No No No Yes
Test coverage 90%+ Unknown Unknown Unknown

We're not saying the other MCP servers are bad. That would be rude, and also legally questionable. They're all built by talented engineers who are contributing to open source and making the ecosystem better. Some of them have features we don't, like the ability to write to your Datadog account from an AI chat window. We chose not to include those. On purpose. You're welcome.

The Philosophy

Read everything. Write nothing.

Your AI assistant should be able to see all of Datadog. It should not be able to silence your pager at 3 AM because it decided your alerts were "probably fine." That's not a hypothetical scenario. That's a Tuesday. We have opinions about this, and they are all correct.


Quick Start```bash

npx github:dreamiurg/datadog-mcp --help


**That's it.** No installation required. Node.js 20+. Works in 8 seconds. We timed it.

---

## Setup

### 1. Get Datadog Credentials

You need two keys from your Datadog Organization Settings:

| Credential | Where to Find | Difficulty |
|------------|---------------|-----------|
| **API Key** | Organization Settings -> API Keys -> New Key | Easy |
| **Application Key** | Organization Settings -> Application Keys -> New Key | Also easy |

> **Tip:** For least-privilege access, scope your Application Key to only the permissions you need. Or give it everything. We won't judge. (We will judge a little.)

### 2. Configure Your AI Tool

Claude Code (CLI)```bash
claude mcp add datadog -- npx github:dreamiurg/datadog-mcp \
  --apiKey YOUR_API_KEY \
  --appKey YOUR_APP_KEY \
  --site datadoghq.com

Or add to ~/.claude.json:```json
{
"mcpServers": {
"datadog": {
"command": "npx",
"args": ["github:dreamiurg/datadog-mcp", "--apiKey", "YOUR_API_KEY", "--appKey", "YOUR_APP_KEY"]
}
}
}


Claude Desktop

Add to your config file:
- **macOS:** `~/Library/Application Support/Claude/claude_desktop_config.json`
- **Windows:** `%APPDATA%/Claude/claude_desktop_config.json````json
{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["github:dreamiurg/datadog-mcp", "--apiKey", "YOUR_API_KEY", "--appKey", "YOUR_APP_KEY"]
    }
  }
}

Cursor / VS Code / Windsurf

Tool Config File
Cursor ~/.cursor/mcp.json or .cursor/mcp.json (project)
Windsurf ~/.codeium/windsurf/mcp_config.json
VS Code User Settings JSON (Ctrl+Shift+P -> "Open User Settings (JSON)")
{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["github:dreamiurg/datadog-mcp", "--apiKey", "YOUR_API_KEY", "--appKey", "YOUR_APP_KEY"]
    }
  }
}

Gemini CLI

Add to ~/.gemini/settings.json:```json
{
"mcpServers": {
"datadog": {
"command": "npx",
"args": ["github:dreamiurg/datadog-mcp", "--apiKey", "YOUR_API_KEY", "--appKey", "YOUR_APP_KEY"]
}
}
}


Codex CLI (OpenAI)

Add to `~/.codex/config.toml`:```toml
[mcp_servers.datadog]
command = "npx"
args = ["github:dreamiurg/datadog-mcp", "--apiKey", "YOUR_API_KEY", "--appKey", "YOUR_APP_KEY"]

The Full Arsenal (117 Tools)

Logs & Log Management (7 tools)

Tool What It Does Try Asking
search-logs Search logs with Datadog query syntax "Find errors in auth service last hour"
aggregate-logs Compute counts, averages, group-bys on log data "Count errors by service in the last 24h"
get-log-pipelines List log pipeline configurations "How are our logs being processed?"
get-log-indexes List log indexes and retention settings "What retention is configured for our logs?"
get-logs-pipeline-order Get log pipeline processing order "What order do log pipelines execute in?"
get-logs-archive-order Get log archive priority order "Which archives are checked first for rehydration?"
list-logs-metrics List log-based custom metrics "What metrics are generated from logs?"

APM & Traces (7 tools)

Tool What It Does Try Asking
search-spans Search APM spans with query syntax "Find slow requests in payment service"
aggregate-spans Compute APM statistics (p99, error rates) "Show p99 latency by service"
get-services List APM-instrumented services "What services are being traced?"
get-trace Get a full distributed trace by ID "Show all spans for trace abc123"
search-apm-events Search raw APM events "Find APM events with errors in checkout"
list-active-apm-events List currently active APM events "What APM events are happening right now?"
list-spans-metrics List span-based APM metrics "What custom span metrics are configured?"

Metrics (6 tools)

Tool What It Does Try Asking
query-metrics Query timeseries metric data "Show CPU usage for web hosts last 4 hours"
get-metrics List available metrics "What metrics are available?"
get-metric-metadata Get metric metadata and descriptions "Describe the system.cpu.user metric"
search-metric-volumes Search metrics by name with volume data "Which metrics matching 'aws.' have the highest volume?"*
get-metric-tag-config Get tag configuration for a specific metric "What tags are indexed for system.cpu.user?"
list-metric-tag-configs List metric tag configurations "Which metrics have custom tag configurations?"

Monitors & Alerting (4 tools)

Tool What It Does Try Asking
get-monitors List monitors (filter by state, tags) "Show all alerting monitors"
get-monitor Get monitor details by ID "Get details for monitor 12345"
search-monitors Search monitors by query "Find monitors related to database"
get-monitor-config-policies Get monitor configuration policies "What monitor config policies are enforced?"

Dashboards (3 tools)

Tool What It Does Try Asking
get-dashboards List all dashboards "What dashboards do we have?"
get-dashboard Get dashboard details and widgets "Show the API metrics dashboard"
list-dashboard-lists List dashboard collections "What dashboard lists are organized?"

SLOs (5 tools)

Tool What It Does Try Asking
get-slos List Service Level Objectives "Which SLOs are breaching?"
get-slo Get SLO details by ID "Show error budget for SLO xyz"
get-slo-history Get historical SLO performance "How has our API SLO trended this month?"
get-slo-corrections List SLO correction periods "What SLO corrections have been applied?"
search-slos Search SLOs with query filters "Find all SLOs with error budget below 10%"

Synthetics (5 tools)

Tool What It Does Try Asking
get-synthetic-tests List synthetic tests "What synthetic checks do we have?"
get-synthetic-results Get results for a specific test "Show recent results for the checkout test"
list-synthetics-global-variables List Synthetics global variables "What global variables are available for tests?"
list-synthetics-locations List available Synthetics testing locations "Where can I run Synthetics tests from?"
list-synthetics-private-locations List Synthetics private locations "What private locations are configured for testing?"

Real User Monitoring (RUM) (3 tools)

Tool What It Does Try Asking
search-rum-events Search RUM events with filters "Find RUM errors on the checkout page"
list-rum-applications List registered RUM applications "What apps have RUM enabled?"
aggregate-rum-events Analytics on RUM data (counts, averages, group-by) "Average page load time by country"

Infrastructure (8 tools)

Tool What It Does Try Asking
get-hosts List infrastructure hosts "Show all production hosts"
get-host-tags Get tags for all hosts "What tags are on our hosts?"
get-active-hosts-count Get total active/up host count "How many hosts are running?"
list-host-totals Get total active and up host counts "Show me the total host count breakdown"
list-network-devices List NDM network devices "Show network device status"
list-fleet-agents List Datadog agents across fleet "Which agents are outdated?"
get-containers List running containers "Show containers filtered by image"
list-processes List running processes "Find java processes across hosts"

Service Catalog (3 tools)

Tool What It Does Try Asking
list-service-definitions List all service definitions "What services are in the catalog?"
get-service-definition Get a single service definition "Who owns the auth service?"
get-service-dependencies Get service dependency map "What does the payment service depend on?"

CI Visibility (4 tools)

Tool What It Does Try Asking
list-ci-pipelines List CI pipeline events "Show recent CI pipeline runs"
get-ci-pipeline-events Aggregate CI pipeline analytics "Average pipeline duration by repo"
list-ci-tests List CI test events "Show failed tests in the last hour"
search-ci-test-events Search CI test events with filters "Find flaky tests in the checkout service"

Security (10 tools)

Tool What It Does Try Asking
search-security-findings Search Cloud Security findings "Find high-severity security findings"
get-security-finding Get a specific security finding "Get finding abc123"
list-posture-findings List compliance/posture findings "Show failed compliance checks"
search-security-signals Search security monitoring signals "Recent threat detections"
get-csm-coverage Get CSM coverage across cloud accounts "Which AWS accounts lack security coverage?"
list-vulnerabilities List vulnerability findings (SAST, SCA, IAST, Infra) "Show open SAST vulnerabilities in production"
list-csm-threats-agent-rules List CSM Threats agent rules "What workload security rules are enabled?"
list-security-rules List security monitoring rules "What detection rules are active?"
list-security-monitoring-rules List security monitoring detection rules "What security detection rules are configured?"
get-sensitive-data-scanner-config Get Sensitive Data Scanner configuration "What sensitive data scanning rules are active?"

Organization (5 tools)

Tool What It Does Try Asking
get-organization Get Datadog organization info "Show organization details and settings"
list-teams List Datadog teams "What teams exist in our org?"
list-team-members List members of a Datadog team "Who is on the platform team?"
list-users List Datadog users "Who has access to Datadog?"
list-authn-mappings List SAML/OIDC authentication mappings "How do IdP groups map to Datadog roles?"

Events, Downtimes & Database Monitoring (6 tools)

Tool What It Does Try Asking
get-events Fetch events in a time range "What events happened today?"
get-downtimes List scheduled maintenance "Any active downtimes?"
list-downtime-schedules List scheduled downtimes (v2) "What downtimes are scheduled?"
get-dbm-samples Get Database Monitoring query samples "Show slow postgres queries"
get-dbm-query-metrics Get Database Monitoring query metrics "Which queries have the highest latency?"
get-ip-ranges Get Datadog IP ranges for firewall config "What IPs should I allowlist for Datadog?"

Audit & Usage (4 tools)

Tool What It Does Try Asking
search-audit-logs Search audit trail with filters "Who changed configs in the last hour?"
get-hourly-usage Get hourly usage metering by product "Show log ingestion usage trends"
get-top-avg-metrics Get top custom metrics by hourly average "Which custom metrics drive cardinality costs?"
get-estimated-cost Get estimated usage cost data "What's our projected Datadog bill this month?"

Notebooks (2 tools)

Tool What It Does Try Asking
list-notebooks List Datadog notebooks with filtering "Find investigation notebooks"
get-notebook Get a specific notebook by ID "Show notebook 12345"

Containers & Processes (2 tools)

Tool What It Does Try Asking
list-containers List containers with tag filtering "Show running containers in us-east-1"
list-processes List processes across hosts "Find Java processes with high memory"

Network Monitoring (2 tools)

Tool What It Does Try Asking
list-network-devices List NDM network devices "Show network device status"
aggregate-network-connections Aggregate network flow analytics "Top network flows by bandwidth"

Security & Compliance (6 tools)

Tool What It Does Try Asking
list-security-rules List security monitoring rules "What detection rules are active?"
search-error-tracking-issues Search error tracking issues "Top errors by volume in production"
get-error-tracking-issue Get error tracking issue details by ID "Show details for issue abc123"
get-csm-coverage Get CSM coverage across cloud accounts "Which AWS accounts lack security coverage?"
list-vulnerabilities List vulnerability findings (SAST, SCA, IAST, Infra) "Show open SAST vulnerabilities in production"
list-csm-threats-agent-rules List CSM Threats agent rules "What workload security rules are enabled?"

Incidents (4 tools)

Tool What It Does Try Asking
search-incidents Search incidents with advanced filters "Find P1 incidents related to database"
get-incident-todos Get action items for an incident "What's pending for this P0?"
get-incident-timeline Get timeline events for an incident "Show the timeline for incident xyz"
get-incident-services List incident services "What services are configured for incident management?"

Cloud & Integrations (7 tools)

Tool What It Does Try Asking
list-aws-accounts List integrated AWS accounts "Which AWS accounts are connected?"
list-gcp-integration List GCP integration accounts "Which GCP projects are integrated?"
list-azure-integration List Azure integration accounts "Which Azure subscriptions are connected?"
list-cloudflare-accounts List Cloudflare accounts "What Cloudflare accounts are integrated?"
list-confluent-accounts List Confluent Cloud accounts "Which Confluent Cloud accounts are connected?"
list-webhooks List webhook integrations "What webhooks are configured?"
list-api-keys List Datadog API keys "What API keys are active?"

DevOps & Automation (6 tools)

Tool What It Does Try Asking
list-dora-deployments List DORA deployment events "Show deployment frequency for web service"
list-workflows List workflow automations "What incident response workflows exist?"
list-workflow-executions List execution instances for a workflow "Show run history for the incident workflow"
list-fleet-agents List Datadog agents across fleet "Which agents are outdated?"
list-monitor-notification-rules List monitor notification routing "Who gets alerted for this monitor?"
list-cost-budgets List cloud cost budgets "Are any teams over budget?"

Access Control (3 tools)

Tool What It Does Try Asking
list-restriction-policies Get restriction policy for a resource "What access restrictions are configured?"
list-app-keys List application keys for the current user "What application keys do I have?"
list-authn-mappings List authentication mappings "How do SAML groups map to Datadog roles?"

The Competitive Landscape (We Did the Research So You Don't Have To)

There are approximately seven thousand Datadog MCP servers on GitHub. We counted. Then we lost count. Then we counted again. Here's how they stack up:

Capability Us Project 1 Project 2 Project 3 Project 4
Total tools 117 ~20 ~10 ~13 ~29
Read-only by design Yes No Yes Yes No
Can mute your hosts No Yes No No No
Can delete your dashboards No No No No Yes
Can schedule downtimes No Yes No No Yes
Test suite 570 tests - - - -
Last meaningful update Today Recent Dec 2025 Jan 2026 Dec 2025

> A note on write operations: Some servers let your AI create monitors, delete dashboards, and schedule downtimes. That's a feature. It's also the plot of every "AI gone wrong" movie ever made. We chose the boring path: read everything, touch nothing. Our therapist says this is healthy.

What They Have That We Don't

In the interest of full transparency (and because our lawyers said we had to):

  • Some servers can mute/unmute hosts and schedule downtimes (write operations we intentionally excluded because we enjoy sleeping)
  • Some servers can CRUD monitors and dashboards (see above re: AI movies, and also re: sleeping)
  • Some servers have unique metric field exploration tools

We sleep extremely well at night.


Configuration Reference

Regional Endpoints

Set --site for your Datadog region:

Region Site Value
US (default) datadoghq.com
EU datadoghq.eu
US3 (GovCloud) ddog-gov.com
US5 us5.datadoghq.com
AP1 ap1.datadoghq.com

Environment Variables

Alternative to command-line arguments:```bash
DD_API_KEY=your_api_key
DD_APP_KEY=your_app_key
DD_SITE=datadoghq.com # Optional, defaults to datadoghq.com
DD_LOGS_SITE=logs.datadoghq.com # Optional, override for logs API
DD_METRICS_SITE=api.datadoghq.com # Optional, override for metrics API


### Logging

| Variable | Values | Default | Description |
|----------|--------|---------|-------------|
| `LOG_LEVEL` | `debug`, `info`, `warn`, `error` | `info` | Minimum log level |
| `LOG_FORMAT` | `json`, `pretty` | `json` | Output format (`pretty` for local dev) |
```bash
# Pretty logs for development
LOG_FORMAT=pretty npx github:dreamiurg/datadog-mcp --apiKey ... --appKey ...

Application Key Scopes (for least-privilege security)

Create an Application Key with only the scopes you need:

Scope Tools Unlocked
monitors_read get-monitors, get-monitor, search-monitors, get-monitor-config-policies, list-monitor-notification-rules
dashboards_read get-dashboards, get-dashboard, list-dashboard-lists
metrics_read get-metrics, get-metric-metadata, query-metrics, search-metric-volumes, get-metric-tag-config, list-metric-tag-configs
events_read get-events
logs_read_data search-logs, aggregate-logs, get-log-pipelines, get-log-indexes, get-logs-pipeline-order, get-logs-archive-order, list-logs-metrics
hosts_read get-hosts, get-host-tags, get-active-hosts-count, list-host-totals
containers_read list-containers, get-containers
monitors_downtime get-downtimes, list-downtime-schedules
slos_read get-slos, get-slo, get-slo-history, get-slo-corrections, search-slos
apm_read search-spans, aggregate-spans, get-trace, search-apm-events, list-active-apm-events, list-spans-metrics
apm_service_catalog_read get-services, get-service-dependencies, list-service-definitions, get-service-definition
rum_read search-rum-events, list-rum-applications, aggregate-rum-events
synthetics_read get-synthetic-tests, get-synthetic-results, list-synthetics-global-variables, list-synthetics-locations, list-synthetics-private-locations
security_monitoring_findings_read search-security-findings, get-security-finding, list-posture-findings, search-security-signals, list-vulnerabilities
security_monitoring_rules_read list-security-rules, list-security-monitoring-rules
csm_agents_read get-csm-coverage, list-csm-threats-agent-rules
teams_read list-teams, list-team-members
user_access_read list-users
notebooks_read list-notebooks, get-notebook
ci_visibility_pipelines_read list-ci-pipelines, get-ci-pipeline-events
ci_visibility_read list-ci-tests, search-ci-test-events
error_tracking_read search-error-tracking-issues, get-error-tracking-issue
audit_trail_read search-audit-logs
dbm_read get-dbm-samples, get-dbm-query-metrics
usage_read get-hourly-usage, get-top-avg-metrics, get-estimated-cost
ndm_read list-network-devices, aggregate-network-connections
aws_configuration_read list-aws-accounts
api_keys_read list-api-keys, list-app-keys
auth_n_mappings_read list-authn-mappings
incident_read search-incidents, get-incident-todos, get-incident-timeline, get-incident-services
dora_deployment_read list-dora-deployments
workflows_read list-workflows, list-workflow-executions
fleet_read list-fleet-agents
cost_management_read list-cost-budgets

Create a scoped key: Organization Settings -> Application Keys -> New Key -> Select scopes


Troubleshooting

403 Forbidden

  1. Verify API key and Application key are correct
  2. Check your Application Key has required scopes
  3. Confirm you're using the correct regional endpoint

Connection Issues```bash

Check MCP server status

claude mcp list

View logs (Claude Desktop - macOS)

tail -f ~/Library/Logs/Claude/mcp*.log

View logs (Claude Desktop - Windows)

Get-Content "$env:APPDATA\Claude\Logs\mcp*.log" -Tail 20 -Wait


Debug Mode```bash
LOG_LEVEL=debug LOG_FORMAT=pretty npx github:dreamiurg/datadog-mcp --apiKey ... --appKey ...

Development```bash

git clone https://github.com/dreamiurg/datadog-mcp.git
cd datadog-mcp
npm install
npm run build
npm test


| Script | Description |
|--------|-------------|
| `npm run build` | Compile TypeScript |
| `npm test` | Run tests (570 tests in 120 files) |
| `npm run test:coverage` | Run tests with coverage (90%+ enforced) |
| `npm run lint` | Run Biome linter |
| `npm run typecheck` | Type check |
| `npm run complexity` | Check cyclomatic complexity |
| `npm run ci` | Run everything (the full gauntlet) |

### Test with MCP Inspector```bash
npx @modelcontextprotocol/inspector node dist/index.js \
  --apiKey=YOUR_API_KEY \
  --appKey=YOUR_APP_KEY

Contributing

PRs welcome. We have strong opinions about read-only access, 90% test coverage, and cognitive complexity under 15. If that sounds like fun, you'll fit right in.


License

MIT - Use freely in personal and commercial projects.


Disclaimer

In case it wasn't abundantly, blindingly, neon-sign-in-the-desert obvious: the comparison tables above are written with tongue so firmly in cheek it's coming out the other side. This is sarcasm. We are being sarcastic. If you're still not sure, yes, that was also sarcasm.

Every single Datadog MCP server out there was built by engineers who cared enough to ship something real into the world. That's more than most people do on a Saturday. Open source is better when there are options, and the ecosystem is better because all of these projects exist.

If our 117 read-only tools aren't what you need, genuinely and sincerely: go find the one that is. Try them all. Mix and match. Build your own. Fork ours and add write operations if you're feeling brave (and have good backups). The best tool is the one that fits your workflow, not the one with the spiciest README.

Now go build something great. Sleep tight.

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
Datadog MCP Server 27315d08 怎么安装?

在客户端按下面的安装命令接入:npx github:dreamiurg/datadog-mcp --help。对应的包名是 github。装完在客户端配置里引用同一个包或端点。

Datadog MCP Server 27315d08 怎么在 Claude Desktop、Claude、Cursor 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Datadog MCP Server 27315d08 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Datadog MCP Server 27315d08 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Datadog MCP Server 27315d08 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、Cursor、Windsurf、VS Code;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Datadog MCP Server 27315d08 能提供哪些能力?

资料显示约 112 个工具;能力标签包括 日志管理、TypeScript、Claude。具体工具清单以官方文档为准。

Datadog MCP Server 27315d08 需要什么运行环境?

运行时为 Node.js;属于本地进程,需要在本机装好运行环境。

Datadog MCP Server 27315d08 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。