clear_dag_run
Clear a DAG run
LABSCO SUMMARY
The operational calls you want during an incident — clear, re-parse, set state, test connection — stop being a browser tab.
The surface follows the REST API closely, so what you can do is exactly what your Airflow account can do, and the destructive calls sit next to the reads with nothing in between. clear_task_instances and clear_dag_run both take dry_run, which is the difference between a plan and an action at three in the morning. Queued dataset events get their own reads and deletes at both levels, which matters only if you schedule on datasets rather than on cron.
What it is
A server over the Airflow REST API, 68 tools covering DAGs, DAG runs, task instances, XComs, datasets, connections, variables, pools and the instance's own health.
What you get
- DAGs fetched by ID pattern or tag, paused and unpaused one at a time or in bulk, their source read, and their files re-parsed after an edit
- DAG runs triggered with a logical date and a note, listed by state or date window, cleared, deleted, or moved to a state by hand
- Task instances cleared with a dry run first, states set across upstream, downstream, past and future selections, and each try listed and read separately
- Connections created, updated and tested before anything depends on them, with variables and pools as first-class objects alongside
- Dataset events read and created by URI, and queued dataset events inspected or dropped at both the DAG and the dataset level
- Instance introspection: health, version, loaded plugins and providers, import errors, and the event log filtered by DAG, task, owner or event type
- XCom entries listed and fetched, deserialised or returned as strings
Requirements
AIRFLOW_HOST pointing at an Airflow deployment reachable from wherever the server runs, plus AIRFLOW_USERNAME and AIRFLOW_PASSWORD (or an AIRFLOW_JWT_TOKEN instead) for an account whose role permits the calls you intend — delete_dag and delete_dag_run sit in the same tool list as the reads.
Setup effort
Add the configuration to your client
USE CASES
01
Restarting a pipeline that died overnight
Find the run that failed, clear the affected task instances with a dry run first to see what the clear would touch, then set the states and move the run itself back to running.
02
Finding out why a DAG stopped appearing
List the import errors, read the one that names your file, check the source Airflow currently holds, and ask it to parse the file again once the fix is in place.
03
Reading the log of the attempt that actually failed
A retried task has more than one log. Read the instance to see how many tries there were, list them, then fetch the log for the try number you care about.
04
Checking a connection before a DAG depends on it
Try the connection details inline first, and only save them as a named connection once the test comes back clean.
05
Answering whether a dataset event ever fired
When a dataset-scheduled DAG has not run, the question is whether the upstream event arrived and whether it is still sitting in the queue.
68 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
clear_task_instances
Clear a set of task instances
create_connection
Create a connection
create_dataset_event
Create dataset event
create_variable
Create a variable
delete_connection
Delete a connection by ID
delete_dag
Delete a DAG
delete_dag_dataset_queued_event
Delete a queued Dataset event for a DAG
delete_dag_dataset_queued_events
Delete queued Dataset events for a DAG
delete_dag_run
Delete a DAG run by DAG ID and DAG run ID
delete_dataset_queued_events
Delete queued Dataset events for a Dataset
delete_pool
Delete a pool
delete_variable
Delete a variable by key
fetch_dags
Fetch all DAGs
get_config
Get current configuration
get_connection
Get a connection by ID
get_dag
Get a DAG by ID
get_dag_dataset_queued_event
Get a queued Dataset event for a DAG
get_dag_dataset_queued_events
Get queued Dataset events for a DAG
get_dag_details
Get a simplified representation of DAG
get_dag_run
Get a DAG run by DAG ID and DAG run ID
get_dag_runs
Get DAG runs by ID
get_dag_runs_batch
List DAG runs (batch)
get_dag_source
Get a source code
get_dag_stats
Get DAG stats
get_dag_tasks
Get tasks for DAG
get_dataset
Get a dataset by URI
get_dataset_events
Get dataset events
get_dataset_queued_events
Get queued Dataset events for a Dataset
get_datasets
List datasets
get_event_log
Get a specific log entry by ID
get_event_logs
List log entries from event log
get_health
Get instance status
get_import_error
Get a specific import error by ID
get_import_errors
List import errors
get_log
Get the log from a task instance by DAG ID, task ID, DAG run ID and task try number
get_plugins
Get a list of loaded plugins
get_pool
Get a pool by name
get_pools
List pools
get_providers
Get a list of loaded providers
get_task
Get a task by ID
get_task_instance
Get a task instance by DAG ID, task ID, and DAG run ID
get_tasks
Get tasks for DAG
get_upstream_dataset_events
Get dataset events for a DAG run
get_value
Get a specific option from configuration
get_variable
Get a variable by key
get_version
Get version information
get_xcom_entries
Get all XCom entries
get_xcom_entry
Get an XCom entry
list_connections
List all connections
list_task_instance_tries
List task instance tries by DAG ID, DAG run ID, and task ID
list_task_instances
List task instances by DAG ID and DAG run ID
list_variables
List all variables
patch_dag
Update a DAG
patch_dags
Update multiple DAGs
patch_pool
Update a pool
pause_dag
Pause a DAG by ID
post_dag_run
Trigger a DAG by ID
post_pool
Create a pool
reparse_dag_file
Request re-parsing of a DAG file
set_dag_run_note
Update the DagRun note
set_task_instances_state
Set a state of task instances
test_connection
Test a connection
unpause_dag
Unpause a DAG by ID
update_connection
Update a connection by ID
update_dag_run_state
Update a DAG run state by DAG ID and DAG run ID
update_task_instance
Update a task instance by DAG ID, DAG run ID, and task ID
update_variable
Update a variable by key
BEFORE YOU INSTALL
What to know first
Tools you get
68 tools when we connected. --read-only (READ_ONLY=true) exposes only read tools, and --apis selects among 15 API groups, so the recommended read-only setup shows far fewer.
Requires
- uvx on PATH
- AIRFLOW_HOST — your Airflow deployment's URL (defaults to http://localhost:8080)
- AIRFLOW_USERNAME and AIRFLOW_PASSWORD, or AIRFLOW_JWT_TOKEN instead
CONTEXT COST
Moderate Standing cost
Rated by the size of the 68 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
MCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"mcp-server-apache-airflow": {
"command": "uvx",
"args": [
"mcp-server-apache-airflow"
],
"env": {
"AIRFLOW_HOST": "https://<YOUR_AIRFLOW_HOST>",
"AIRFLOW_USERNAME": "<YOUR_AIRFLOW_USERNAME>",
"AIRFLOW_PASSWORD": "<YOUR_AIRFLOW_PASSWORD>"
}
}
}
}This block signs in with a username and password. To use a JWT token instead, put "AIRFLOW_JWT_TOKEN": "<YOUR_AIRFLOW_JWT_TOKEN>" in env instead of AIRFLOW_USERNAME and AIRFLOW_PASSWORD; if both are set, the token takes precedence. AIRFLOW_HOST defaults to http://localhost:8080 when left out.
