agent-observability-mcp
MCP server for AI agent observability, providing trace and span logging, search, latency/tokens/cost metrics, and anomaly detection using an in-memory buffer.
README
agent-observability-mcp
Serveur MCP (Model Context Protocol) d'observabilite pour agents IA : journalisation des traces et des spans, recherche, metriques de latence, de tokens et de cout, detection d'anomalies.
Aucune base de donnees, aucune cle d'API, aucun service tiers : tout tient dans un tampon memoire configurable.
Outils exposes
| Outil | Description |
|---|---|
obs_event_log |
Journalise un span d'agent : operation, trace, statut, duree, modele, tokens, cout, attributs libres. |
obs_trace_get |
Reconstitue la chronologie complete d'une trace : spans ordonnes, duree cumulee, erreurs, tokens, cout. |
obs_events_search |
Recherche des evenements par projet, operation, type, statut, modele, fenetre temporelle, duree minimale ou texte libre. |
obs_metrics_summary |
Agrege volume, taux d'erreur, latences p50 / p95 / p99, tokens, cout et top 5 des operations. |
obs_anomaly_scan |
Signale taux d'erreur anormaux, latences aberrantes, concentration de cout et boucles de retry. |
obs_projects_list |
Liste les projets instrumentes et le remplissage du tampon. |
Chaque outil accepte response_format : markdown ou json.
Modele de donnees
Un evenement represente un span : trace_id, span_id, parent_span_id, name, kind (llm, tool, retrieval, agent, http, other), status (ok, error, timeout), duration_ms, model, input_tokens, output_tokens, cost_usd, attributes.
Installation locale
npm install
npm run build
node dist/index.js
Connect via MCPize
Use this MCP server instantly with no local installation:
npx -y mcpize connect @contact.agentia1984/agent-observability --client claude
Or connect at: https://mcpize.com/mcp/agent-observability
Configuration Claude Desktop
{
"mcpServers": {
"agent-observability": {
"command": "node",
"args": ["/chemin/vers/agent-observability-mcp/dist/index.js"]
}
}
}
Deploiement heberge (HTTP Streamable)
TRANSPORT=http node dist/index.js
Le serveur ecoute sur 0.0.0.0 et lit process.env.PORT (8083 par defaut). Endpoints : POST /mcp, POST / et GET /health.
Variables d'environnement
| Variable | Valeur | Role |
|---|---|---|
TRANSPORT |
stdio (defaut) ou http |
Choix du transport MCP. |
PORT |
entier | Port d'ecoute en mode HTTP. |
OBS_MAX_EVENTS |
entier, defaut 5000 | Taille du tampon circulaire d'evenements. |
Limites connues
Le stockage est en memoire : les evenements sont perdus au redemarrage et ne sont pas partages entre plusieurs instances. Pour un usage longue duree, journalisez en parallele vers votre propre entrepot.
Licence
MIT.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.