jaeger-mcp
Provides read-only access to Jaeger distributed tracing data, enabling search for traces, inspection of spans, and mapping of service dependencies through the Jaeger API.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@jaeger-mcpshow me traces with errors in the payment service from the last hour"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
jaeger-mcp
MCP server for Jaeger distributed tracing. Give Claude (or any MCP-capable agent) read access to your trace data — search traces, inspect spans, map service dependencies — without leaving the conversation.
Why another Jaeger MCP?
The existing Jaeger integrations require a running UI or custom scripts. This server:
Speaks the standard Model Context Protocol over stdio — works with Claude Desktop, Claude Code, Cursor, and any MCP client.
Is read-only: all 5 tools carry
readOnlyHint: true— zero risk of modifying trace data.Returns dual-channel output: structured JSON (
structuredContent) for programmatic use + Markdown (content) for human-readable display.Has actionable error messages that name the exact env var to fix and suggest a next step.
Supports Bearer token, HTTP Basic auth, or no auth (common for internal deployments).
Tools
Tool | Endpoint | Description |
|
| List all instrumented services |
|
| List operation names for a service |
|
| Search traces with rich filters |
|
| Full trace detail with span tree |
|
| Service-to-service call graph |
Installation
pip install jaeger-mcpOr run directly without installing:
uvx jaeger-mcpConfiguration
All configuration is via environment variables:
Variable | Required | Default | Description |
| Yes | — | Jaeger query service URL, e.g. |
| No | — | Bearer token (takes precedence over Basic auth) |
| No | — | HTTP Basic auth username |
| No | — | HTTP Basic auth password |
| No |
| Set |
Copy .env.example to .env and fill in your values.
Claude Desktop / Claude Code setup
Add to your MCP config (claude_desktop_config.json or .claude/mcp.json):
{
"mcpServers": {
"jaeger": {
"command": "jaeger-mcp",
"env": {
"JAEGER_URL": "https://jaeger.example.com",
"JAEGER_TOKEN": "your-token-here"
}
}
}
}Or with uvx (no install required):
{
"mcpServers": {
"jaeger": {
"command": "uvx",
"args": ["jaeger-mcp"],
"env": {
"JAEGER_URL": "https://jaeger.example.com"
}
}
}
}Docker
docker run --rm -e JAEGER_URL=https://jaeger.example.com jaeger-mcpExample queries
Once configured, ask Claude:
"What services does Jaeger know about?"
"Find traces with HTTP 500 errors in
order-servicefrom the last hour""Show me the slowest traces (over 2 seconds) for
GET /checkout""What caused the error in trace
abcdef1234567890?""Map the service dependency graph for the last 7 days"
"Which services call
postgresmost frequently?"
Tool usage guide
jaeger_list_services
Returns all service names Jaeger has seen. Start here when you don't know which services are instrumented. Output is capped at 500 services with a truncation hint.
jaeger_list_operations
Returns all operation names for a given service (e.g. HTTP route names, gRPC method names). Use to discover valid operation names before filtering jaeger_search_traces.
jaeger_search_traces
The main search tool. Filters:
service(required) — service name fromjaeger_list_servicesoperation— narrow to a specific endpointtags— JSON string of tag filters, e.g.{"http.status_code":"500"}or{"error":"true"}start/end— time range in microseconds UTCmin_duration/max_duration— duration strings like"100ms","1.5s","2m"limit— default 20, max 1500
Returns trace summaries with trace_id, duration_us, span_count, service_count, root_operation, errors_count.
jaeger_get_trace
Full trace detail. Accepts a trace_id (hex string, 16-32 chars) and returns:
All spans with tags, service names, parent/child relationships
Per-service statistics (span count, total duration, error count)
Execution tree (each node lists its child span IDs)
Error spans are identified by tags["error"] = "true".
jaeger_get_dependencies
Service topology graph. Returns directed edges (parent → child) with call_count. Use lookback_hours (default 24, max 720) to control the window.
Performance characteristics
All tools use a single persistent
requests.Sessionwith connection pooling.The session has
trust_env = Falseto bypass environment proxies (Jaeger is typically an internal service).Requests time out after 30 seconds.
jaeger_search_tracespasseslimitdirectly to Jaeger — avoid requesting more traces than needed.jaeger_get_tracefetches the full trace in one call — large traces (thousands of spans) may be slow.jaeger_get_dependenciesaggregates over the full lookback window; large windows may be slow on busy clusters.
Development
git clone https://github.com/mshegolev/jaeger-mcp
cd jaeger-mcp
pip install -e '.[dev]'
pytest tests/ -v
ruff check src tests
ruff format src testsLicense
MIT — see LICENSE.
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/mshegolev/jaeger-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server