Servidor MCP de Airflow: lee DAGs, ejecuciones, instancias de tareas, logs; dispara y limpia (protegido contra escritura). Construido sobre
@us-all/mcp-toolkit.
Un MCP enfocado en la API REST estable de Airflow. Lectura por defecto; airflow-trigger-dag y airflow-clear-task están protegidos detrás de AIRFLOW_ALLOW_WRITE=true.
Para una integración más profunda con dbt (análisis de manifiestos, historial de resultados de ejecución, frescura de fuentes, cobertura de pruebas por columna, recorridos de linaje, tablas de resultados de DQ personalizadas), instala el compañero @us-all/dbt-mcp junto a este.
7 herramientas en las categorías airflow + meta
2 Prompts de MCP para flujos de trabajo de triaje de DAGs
1 herramienta de agregación que reemplaza la combinación list-runs → get-task-instances
pnpm add -D @us-all/airflow-mcpRelated MCP server: Apache Airflow MCP Server
AIRFLOW_API_URL=http://airflow.example.com:8080/api/v1 \
AIRFLOW_USERNAME=admin AIRFLOW_PASSWORD=... \
npx @us-all/airflow-mcpEl servidor utiliza MCP stdio; conéctalo a Claude Desktop / Cursor / cualquier cliente MCP. Configura MCP_TRANSPORT=http para optar por el transporte HTTP transmitible (autenticación Bearer, endpoint /health).
airflow (6 + 1 agregación)Herramienta | Descripción |
| Lista DAGs activos con filtros de etiqueta/búsqueda |
| Ejecuciones recientes de un DAG (filtro de estado, ordenadas de más reciente a más antigua) |
| Instancias de tarea para una ejecución de DAG específica |
| Tail (últimos N kB) del log de una instancia de tarea |
| Dispara una nueva ejecución (protegido contra escritura) |
| Limpia instancias de tarea específicas → re-ejecuta (protegido contra escritura) |
| Salud agregada del DAG: tasa de éxito + duración promedio + última ejecución fallida + tareas fallidas |
metasearch-tools — descubrimiento de herramientas mediante lenguaje natural.
Prompt | Usar cuando |
| "¿Por qué falló el DAG X?" — extrae ejecuciones, aísla el fallo, vuelca logs, propone remediación |
| "Revisar todos los DAGs en busca de baja tasa de éxito y ejecuciones bloqueadas" |
Env | Requerido | Notas |
| sí | Base de la API REST de Airflow, ej. |
| no | Nombre de usuario para autenticación básica |
| no | Contraseña para autenticación básica (secreto) |
| no |
|
| no | Interruptores de categoría |
API REST estable de Airflow (2.x). Airflow 3.x expone la misma superficie a través de los endpoints estables /api/v1/, por lo que este servidor funciona tanto con despliegues 2.x como 3.x.
Para el análisis de artefactos de dbt, historial de resultados de ejecución y tablas de resultados de DQ, instala @us-all/dbt-mcp junto a este.
pnpm install
pnpm run build # tsc → dist/
pnpm test # vitest
pnpm run smoke # spawns dist/index.js, calls initialize + tools/list (set env first)MIT — ver LICENSE.
Clear specific task instances in an Airflow DAG run (re-run them); supports include_upstream / include_downstream. Write-gated by AIRFLOW_ALLOW_WRITE.
Inputs: dagId (required), dagRunId (required), taskIds (required), includeDownstream, +1 more.
List task instances for a specific Airflow DAG run with state, try_number, duration
Inputs: dagId (required), dagRunId (required), extractFields.
Fetch the tail (last N kB) of an Airflow task instance log for a specific try_number
Inputs: dagId (required), dagRunId (required), taskId (required), tryNumber, +2 more.
List Airflow DAGs (active by default) with optional tag filter and dag_id substring search
Inputs: onlyActive, tag, search, limit, +1 more.
List recent runs of one Airflow DAG, optionally filtered by state, ordered newest first
Inputs: dagId (required), state, limit, extractFields.
Trigger a new Airflow DAG run with optional conf payload and note. Write-gated by AIRFLOW_ALLOW_WRITE.
Inputs: dagId (required), dagRunId, conf, note, +1 more.
Aggregated DAG health: success-rate over the last N runs + count breakdown (succeeded/failed/queued) + average duration + last-failed-run id + (optional) failing task instances. Replaces the airflow-list-runs + airflow-get-task-instances combo for 'is this DAG healthy right now?'.
Inputs: dagId (required), recentRuns, includeFailingTasks, extractFields.
Discover available Airflow MCP tools by natural language query.
Inputs: query (required), category, limit.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 8 tools
Each tool maps to a distinct resource/action: DAGs, runs, task instances, logs, aggregate health, and tool discovery. dag-health-rollup and airflow-list-runs could overlap for health questions, but the description explicitly positions the rollup as the higher-level replacement, so ambiguity remains low.
Most tools follow the airflow-<verb>-<object> pattern (airflow-list-dags, airflow-trigger-dag, airflow-clear-task, airflow-list-runs, airflow-get-task-instances, airflow-get-task-logs), but dag-health-rollup and search-tools break both the airflow- prefix and the verb-noun convention. There is also minor inconsistency between 'list' and 'get' for similar retrieval operations.
Eight tools is well-scoped for an Airflow operations server. Each tool covers a meaningful operation without redundancy or bloat.
The set covers the core operational workflow: discover DAGs, trigger runs, inspect runs/task instances/logs, clear tasks, and get an aggregated health rollup. Minor gaps like pause/unpause DAGs or fetching DAG details exist, but they do not break the primary monitoring and troubleshooting use case.
Official Airtable MCP server — database and operations layer for agents.
Streamable HTTP MCP server exposing planner flows, tasks, and squads.
MCP server for mandates, delegation, policy-gated execution, credential grants, and audit.
Argentic MCP server: run and inspect long-running agents on Argentic.
