datadog-logs-mcp
# datadog-logs-mcp
MCP server for querying Datadog logs. Use with Claude Code, Claude Desktop, or any MCP-compatible client.
## Setup
```json
{
"mcpServers": {
"datadog-logs": {
"command": "npx",
"args": ["-y", "datadog-logs-mcp"],
"env": {
"DD_API_KEY": "your-api-key",
"DD_APP_KEY": "your-app-key",
"DD_SITE": "datadoghq.com"
}
}
}
}
```
### Environment Variables
| Variable | Required | Description |
|----------|----------|-------------|
| `DD_API_KEY` | Yes | Datadog API key |
| `DD_APP_KEY` | Yes | Datadog Application key |
| `DD_SITE` | No | Datadog site (default: `datadoghq.com`). Use `datadoghq.eu`, `us3.datadoghq.com`, `us5.datadoghq.com`, etc. |
## Tools
### `search_logs`
Search logs using Datadog log search syntax with time range, sorting, and pagination.
### `get_log`
Get a specific log entry by ID.
### `aggregate_logs`
Aggregate logs with computations (count, avg, sum, min, max, percentiles) and group-by breakdowns.
### `list_indexes`
List all configured log indexes.
## License
MIT
TDQS
Scored across 3 tools
Each tool targets a distinct operation: search_logs filters and returns multiple entries, get_log fetches a single entry by ID, and aggregate_logs computes metrics. There is no meaningful overlap, making tool selection unambiguous.
All tool names follow the same verb_noun snake_case pattern (search_logs, get_log, aggregate_logs), which is predictable and consistent across the set.
Three tools is a well-scoped count for a log-focused server. It covers the core operations (search, retrieve, aggregate) without unnecessary bloat or thinness.
The set covers the primary log reading and analysis workflow: searching, fetching by ID, and aggregating metrics. Minor gaps like log index listing or log ingestion are absent but not essential for the apparent purpose.