Managing Connections¶

Storing a Connection in Environment Variables¶

Airflow connections may be defined in environment variables.

The naming convention is AIRFLOW_CONN_{CONN_ID}, all uppercase (note the single underscores surrounding CONN). So if your connection id is my_prod_db then the variable name should be AIRFLOW_CONN_MY_PROD_DB.

The value can be either JSON or Airflow’s URI format.

JSON example¶

New in version 2.3.0.

If serializing with JSON:

export AIRFLOW_CONN_MY_PROD_DATABASE='{
    "conn_type": "my-conn-type",
    "login": "my-login",
    "password": "my-password",
    "host": "my-host",
    "port": 1234,
    "schema": "my-schema",
    "extra": {
        "param1": "val1",
        "param2": "val2"
    }
}'

URI example¶

If serializing with Airflow URI:

export AIRFLOW_CONN_MY_PROD_DATABASE='my-conn-type://login:password@host:port/schema?param1=val1&param2=val2'

See Connection URI format for more details on how to generate the a valid URI.

Note

Connections defined in environment variables will not show up in the Airflow UI or using airflow connections list.

Storing connections in a Secrets Backend¶

You can store Airflow connections in external secrets backends like HashiCorp Vault, AWS SSM Parameter Store, and other such services. For more details see Secrets Backend.

Storing connections in the database¶

Creating a Connection with the UI¶

Open the Admin->Connections section of the UI. Click the Create link to create a new connection.

Fill in the Connection Id field with the desired connection ID. It is recommended that you use lower-case characters and separate words with underscores.
Choose the connection type with the Connection Type field.
Fill in the remaining fields. See Handling of arbitrary dict in extra for a description of the fields belonging to the different connection types.
Click the Save button to create the connection.

Editing a Connection with the UI¶

Open the Admin->Connections section of the UI. Click the pencil icon next to the connection you wish to edit in the connection list.

Modify the connection properties and click the Save button to save your changes.

Creating a Connection from the CLI¶

You may add a connection to the database from the CLI.

You can add a connection using JSON format (from version 2.3.0):

airflow connections add 'my_prod_db' \
    --conn-json '{
        "conn_type": "my-conn-type",
        "login": "my-login",
        "password": "my-password",
        "host": "my-host",
        "port": 1234,
        "schema": "my-schema",
        "extra": {
            "param1": "val1",
            "param2": "val2"
        }
    }'

Alternatively you may use Airflow’ Connection URI format (see Generating a Connection URI).

airflow connections add 'my_prod_db' \
    --conn-uri '<conn-type>://<login>:<password>@<host>:<port>/<schema>?param1=val1&param2=val2&...'

Lastly, you may also specify each parameter individually:

airflow connections add 'my_prod_db' \
    --conn-type 'my-conn-type' \
    --conn-login 'login' \
    --conn-password 'password' \
    --conn-host 'host' \
    --conn-port 'port' \
    --conn-schema 'schema' \
    ...

Exporting connections to file¶

You can export to file connections stored in the database (e.g. for migrating connections from one environment to another). See Exporting Connections for usage.

Security of connections in the database¶

For connections stored in the Airflow metadata database, Airflow uses Fernet to encrypt password and other potentially sensitive data. It guarantees that without the encryption password, Connection Passwords cannot be manipulated or read without the key. For information on configuring Fernet, look at Fernet.

Testing Connections¶

Airflow Web UI & API allows to test connections. The test connection feature can be used from create or edit connection page, or through calling Connections REST API.

To test a connection Airflow calls out the test_connection method from the associated hook class and reports the results of it. It may happen that the connection type does not have any associated hook or the hook doesn’t have the test_connection method implementation, in either case the error message will throw the proper error message.

One important point to note is that the connections will be tested from the webserver only, so this feature is subject to network egress rules setup for your webserver. Also, if webserver & worker machines have different libs or provider packages installed then the test results might differ.

Last caveat is that this feature won’t be available for the connections coming out of the secrets backends.

Custom connection types¶

Airflow allows the definition of custom connection types – including modifications of the add/edit form for the connections. Custom connection types are defined in community maintained providers, but you can can also add a custom provider that adds custom connection types. See Provider packages for description on how to add custom providers.

The custom connection types are defined via Hooks delivered by the providers. The Hooks can implement methods defined in the protocol class DiscoverableHook. Note that your custom Hook should not derive from this class, this class is a dummy example to document expectations regarding about class fields and methods that your Hook might define. Another good example is JdbcHook.

By implementing those methods in your hooks and exposing them via connection-types array (and deprecated hook-class-names) in the provider meta-data, you can customize Airflow by:

Adding custom connection types
Adding automated Hook creation from the connection type
Adding custom form widget to display and edit custom “extra” parameters in your connection URL
Hiding fields that are not used for your connection
Adding placeholders showing examples of how fields should be formatted

You can read more about details how to add custom provider packages in the Provider packages

Note

Deprecated hook-class-names

Prior to Airflow 2.2.0, the connections in providers have been exposed via hook-class-names array in provider’s meta-data, this however has proven to be not well optimized for using individual hooks in workers and the hook-class-names array is now replaced by connection-types array. Until provider supports Airflow below 2.2.0, both connection-types and hook-class-names should be present. Automated checks during CI build will verify consistency of those two arrays.

URI format¶

Note

From version 2.3.0 you can serialize connections with JSON instead. See example.

For historical reasons, Airflow has a special URI format that can be used for serializing a Connection object to a string value.

In general, Airflow’s URI format looks like the following:

my-conn-type://my-login:my-password@my-host:5432/my-schema?param1=val1&param2=val2

The above URI would produce a Connection object equivalent to the following:

Connection(
    conn_id="",
    conn_type="my_conn_type",
    description=None,
    login="my-login",
    password="my-password",
    host="my-host",
    port=5432,
    schema="my-schema",
    extra=json.dumps(dict(param1="val1", param2="val2")),
)

Generating a connection URI¶

To make connection URI generation easier, the Connection class has a convenience method get_uri(). It can be used like so:

>>> import json
>>> from airflow.models.connection import Connection
>>> c = Connection(
...     conn_id="some_conn",
...     conn_type="mysql",
...     description="connection description",
...     host="myhost.com",
...     login="myname",
...     password="mypassword",
...     extra=json.dumps(dict(this_param="some val", that_param="other val*")),
... )
>>> print(f"AIRFLOW_CONN_{c.conn_id.upper()}='{c.get_uri()}'")
AIRFLOW_CONN_SOME_CONN='mysql://myname:mypassword@myhost.com?this_param=some+val&that_param=other+val%2A'

Additionally, if you have created a connection, you can use airflow connections get command.

$ airflow connections get sqlite_default
Id: 40
Connection Id: sqlite_default
Connection Type: sqlite
Host: /tmp/sqlite_default.db
Schema: null
Login: null
Password: null
Port: null
Is Encrypted: false
Is Extra Encrypted: false
Extra: {}
URI: sqlite://%2Ftmp%2Fsqlite_default.db

Handling of arbitrary dict in extra¶

Some JSON structures cannot be urlencoded without loss. For such JSON, get_uri will store the entire string under the url query param __extra__.

For example:

>>> extra_dict = {"my_val": ["list", "of", "values"], "extra": {"nested": {"json": "val"}}}
>>> c = Connection(
...     conn_type="scheme",
...     host="host/location",
...     schema="schema",
...     login="user",
...     password="password",
...     port=1234,
...     extra=json.dumps(extra_dict),
... )
>>> uri = c.get_uri()
>>> uri
'scheme://user:password@host%2Flocation:1234/schema?__extra__=%7B%22my_val%22%3A+%5B%22list%22%2C+%22of%22%2C+%22values%22%5D%2C+%22extra%22%3A+%7B%22nested%22%3A+%7B%22json%22%3A+%22val%22%7D%7D%7D'

And we can verify that it returns the same dictionary:

>>> new_c = Connection(uri=uri)
>>> new_c.extra_dejson == extra_dict
True

But for the most common case of storing only key-value pairs, plain url encoding is used.

You can verify a URI is parsed correctly like so:

>>> from airflow.models.connection import Connection

>>> c = Connection(
...     uri="my-conn-type://my-login:my-password@my-host:5432/my-schema?param1=val1&param2=val2"
... )
>>> print(c.login)
my-login
>>> print(c.password)
my-password

Handling of special characters in connection params¶

Note

Use the convenience method Connection.get_uri when generating a connection as described in section Generating a Connection URI. This section for informational purposes only.

Special handling is required for certain characters when building a URI manually.

For example if your password has a /, this fails:

>>> c = Connection(
...     uri="my-conn-type://my-login:my-pa/ssword@my-host:5432/my-schema?param1=val1&param2=val2"
... )
ValueError: invalid literal for int() with base 10: 'my-pa'

To fix this, you can encode with quote_plus():

>>> c = Connection(
...     uri="my-conn-type://my-login:my-pa%2Fssword@my-host:5432/my-schema?param1=val1&param2=val2"
... )
>>> print(c.password)
my-pa/ssword