Tempo MCP Server Configuration

The TalkOps Tempo MCP Server supports single-backend and multi-backend configurations, multi-tenancy via X-Scope-OrgID, and strict TraceQL query guardrails.

---

Server & Transport Settings

Variable	Default	Allowed Values	Description
MCP_TRANSPORT	stdio	stdio, http, sse, streamable-http	Communication protocol for MCP clients.
MCP_HOST	0.0.0.0	Valid IP Address	Host interface to bind for HTTP/SSE server.
MCP_PORT	8768	Valid TCP Port	Listening port for HTTP/SSE server.
MCP_PATH	/mcp	URL Path	Base endpoint path for JSON-RPC endpoints.
MCP_LOG_LEVEL	INFO	DEBUG, INFO, WARNING, ERROR	Server logging verbosity.
MCP_LOG_FORMAT	json	json, text	Structured JSON or text log formatting.
MCP_HTTP_TIMEOUT	300	Integer (Seconds)	Maximum client request duration before timing out.

---

Single Backend Configuration

Variable	Default	Description
TEMPO_BASE_URL	http://localhost:3200	Grafana Tempo HTTP API base URL.
TEMPO_BACKEND_ID	default	Identifier for this backend instance.
TEMPO_DISPLAY_NAME	(empty)	Human-readable backend name.
TEMPO_TIMEOUT	30	Request timeout per query in seconds.
TEMPO_VERIFY_SSL	true	Verify SSL certificates for remote endpoints.
TEMPO_AUTH_HEADER	(empty)	Optional authorization header (e.g. Bearer <token>).

---

Multi-Tenancy & Headers

Variable	Default	Description
TEMPO_MULTI_TENANT	false	Enable multi-tenant header injection.
TEMPO_DEFAULT_TENANT	(empty)	Default tenant identifier.
TEMPO_TENANT_HEADER	X-Scope-OrgID	HTTP header name used for tenant identification.

---

Query Policies & Guardrails

Variable	Default	Description
TEMPO_MAX_LOOKBACK	168h (7 days)	Maximum query lookback window.
TEMPO_DEFAULT_SEARCH_LIMIT	20	Default maximum traces returned per search.
TEMPO_MAX_SEARCH_LIMIT	100	Upper limit on traces returned per query.
TEMPO_DEFAULT_SPSS	3	Default spans per span-set.
TEMPO_MAX_SPSS	10	Maximum spans per span-set.
TEMPO_LLM_FORMAT	true	Enable LLM-optimized trace format (Tempo 2.9+ application/vnd.grafana.llm).
TEMPO_MAX_METRICS_DURATION	3h	Maximum allowed TraceQL metrics query time range.

---

Kubernetes Deployment Example

apiVersion: apps/v1
kind: Deployment
metadata:
  name: tempo-mcp-server
  namespace: monitoring
  labels:
    app.kubernetes.io/name: tempo-mcp-server
    app.kubernetes.io/part-of: talkops
spec:
  replicas: 1
  selector:
    matchLabels:
      app: tempo-mcp-server
  template:
    metadata:
      labels:
        app: tempo-mcp-server
    spec:
      containers:
        - name: server
          image: talkopsai/tempo-mcp-server:latest
          imagePullPolicy: IfNotPresent
          env:
            - name: MCP_TRANSPORT
              value: "http"
            - name: MCP_PORT
              value: "8768"
            - name: TEMPO_BASE_URL
              value: "http://tempo-query-frontend.monitoring:3200"
            - name: TEMPO_MULTI_TENANT
              value: "true"
            - name: TEMPO_DEFAULT_TENANT
              value: "production"
          ports:
            - containerPort: 8768
              name: http-mcp
          resources:
            requests:
              cpu: 100m
              memory: 128Mi
            limits:
              cpu: 500m
              memory: 512Mi