Skip to content

Parameters

The parameters utility provides high-level functions to retrieve one or multiple parameter values from AWS Systems Manager Parameter Store, AWS Secrets Manager, AWS AppConfig, Amazon DynamoDB, or bring your own.

Key features

  • Retrieve one or multiple parameters from the underlying provider
  • Cache parameter values for a given amount of time (defaults to 5 seconds)
  • Transform parameter values from JSON or base 64 encoded strings
  • Bring Your Own Parameter Store Provider

Getting started

By default, we fetch parameters from System Manager Parameter Store, secrets from Secrets Manager, and application configuration from AppConfig.

IAM Permissions

This utility requires additional permissions to work as expected.

Note

Different parameter providers require different permissions.

Provider Function/Method IAM Permission
SSM Parameter Store get_parameter, SSMProvider.get ssm:GetParameter
SSM Parameter Store get_parameters, SSMProvider.get_multiple ssm:GetParametersByPath
Secrets Manager get_secret, SecretsManager.get secretsmanager:GetSecretValue
DynamoDB DynamoDBProvider.get dynamodb:GetItem
DynamoDB DynamoDBProvider.get_multiple dynamodb:Query
App Config AppConfigProvider.get_app_config, get_app_config appconfig:GetConfiguration

Fetching parameters

You can retrieve a single parameter using get_parameter high-level function.

For multiple parameters, you can use get_parameters and pass a path to retrieve them recursively.

Fetching multiple parameters recursively
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    # Retrieve a single parameter
    value = parameters.get_parameter("/my/parameter")

    # Retrieve multiple parameters from a path prefix recursively
    # This returns a dict with the parameter name as key
    values = parameters.get_parameters("/my/path/prefix")
    for k, v in values.items():
        print(f"{k}: {v}")

Fetching secrets

You can fetch secrets stored in Secrets Manager using get_secrets.

Fetching secrets
1
2
3
4
5
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    # Retrieve a single secret
    value = parameters.get_secret("my-secret")

Fetching app configurations

You can fetch application configurations in AWS AppConfig using get_app_config.

The following will retrieve the latest version and store it in the cache.

Fetching latest config from AppConfig
1
2
3
4
5
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    # Retrieve a single configuration, latest version
    value: bytes = parameters.get_app_config(name="my_configuration", environment="my_env", application="my_app")

Advanced

Adjusting cache TTL

Tip

max_age parameter is also available in high level functions like get_parameter, get_secret, etc.

By default, we cache parameters retrieved in-memory for 5 seconds.

You can adjust how long we should keep values in cache by using the param max_age, when using get() or get_multiple() methods across all providers.

Caching parameter(s) value in memory for longer than 5 seconds
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

config = Config(region_name="us-west-1")
ssm_provider = parameters.SSMProvider(config=config)

def handler(event, context):
    # Retrieve a single parameter
    value = ssm_provider.get("/my/parameter", max_age=60) # 1 minute

    # Retrieve multiple parameters from a path prefix
    values = ssm_provider.get_multiple("/my/path/prefix", max_age=60)
    for k, v in values.items():
        print(f"{k}: {v}")

Always fetching the latest

If you'd like to always ensure you fetch the latest parameter from the store regardless if already available in cache, use force_fetch param.

Forcefully fetching the latest parameter whether TTL has expired or not
1
2
3
4
5
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    # Retrieve a single parameter
    value = parameters.get_parameter("/my/parameter", force_fetch=True)

Built-in provider class

For greater flexibility such as configuring the underlying SDK client used by built-in providers, you can use their respective Provider Classes directly.

Tip

This can be used to retrieve values from other regions, change the retry behavior, etc.

SSMProvider

Example with SSMProvider for further extensibility
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

config = Config(region_name="us-west-1")
ssm_provider = parameters.SSMProvider(config=config) # or boto3_session=boto3.Session()

def handler(event, context):
    # Retrieve a single parameter
    value = ssm_provider.get("/my/parameter")

    # Retrieve multiple parameters from a path prefix
    values = ssm_provider.get_multiple("/my/path/prefix")
    for k, v in values.items():
        print(f"{k}: {v}")

The AWS Systems Manager Parameter Store provider supports two additional arguments for the get() and get_multiple() methods:

Parameter Default Description
decrypt False Will automatically decrypt the parameter.
recursive True For get_multiple() only, will fetch all parameter values recursively based on a path prefix.
Example with get() and get_multiple()
1
2
3
4
5
6
7
8
from aws_lambda_powertools.utilities import parameters

ssm_provider = parameters.SSMProvider()

def handler(event, context):
    decrypted_value = ssm_provider.get("/my/encrypted/parameter", decrypt=True)

    no_recursive_values = ssm_provider.get_multiple("/my/path/prefix", recursive=False)

SecretsProvider

Example with SecretsProvider for further extensibility
1
2
3
4
5
6
7
8
9
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

config = Config(region_name="us-west-1")
secrets_provider = parameters.SecretsProvider(config=config)

def handler(event, context):
    # Retrieve a single secret
    value = secrets_provider.get("my-secret")

DynamoDBProvider

The DynamoDB Provider does not have any high-level functions, as it needs to know the name of the DynamoDB table containing the parameters.

DynamoDB table structure for single parameters

For single parameters, you must use id as the partition key for that table.

Example

DynamoDB table with id partition key and value as attribute

id value
my-parameter my-value

With this table, dynamodb_provider.get("my-param") will return my-value.

1
2
3
4
5
6
7
from aws_lambda_powertools.utilities import parameters

dynamodb_provider = parameters.DynamoDBProvider(table_name="my-table")

def handler(event, context):
    # Retrieve a value from DynamoDB
    value = dynamodb_provider.get("my-parameter")

You can initialize the DynamoDB provider pointing to DynamoDB Local using endpoint_url parameter:

1
2
3
from aws_lambda_powertools.utilities import parameters

dynamodb_provider = parameters.DynamoDBProvider(table_name="my-table", endpoint_url="http://localhost:8000")

DynamoDB table structure for multiple values parameters

You can retrieve multiple parameters sharing the same id by having a sort key named sk.

Example

DynamoDB table with id primary key, sk as sort keyandvalue` as attribute

id sk value
my-hash-key param-a my-value-a
my-hash-key param-b my-value-b
my-hash-key param-c my-value-c

With this table, dynamodb_provider.get_multiple("my-hash-key") will return a dictionary response in the shape of sk:value.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
from aws_lambda_powertools.utilities import parameters

dynamodb_provider = parameters.DynamoDBProvider(table_name="my-table")

def handler(event, context):
    # Retrieve multiple values by performing a Query on the DynamoDB table
    # This returns a dict with the sort key attribute as dict key.
    parameters = dynamodb_provider.get_multiple("my-hash-key")
    for k, v in parameters.items():
        # k: param-a
        # v: "my-value-a"
        print(f"{k}: {v}")

1
2
3
4
5
{
    "param-a": "my-value-a",
    "param-b": "my-value-b",
    "param-c": "my-value-c"
}

Customizing DynamoDBProvider

DynamoDB provider can be customized at initialization to match your table structure:

Parameter Mandatory Default Description
table_name Yes (N/A) Name of the DynamoDB table containing the parameter values.
key_attr No id Hash key for the DynamoDB table.
sort_attr No sk Range key for the DynamoDB table. You don't need to set this if you don't use the get_multiple() method.
value_attr No value Name of the attribute containing the parameter value.
Customizing DynamoDBProvider to suit your table design
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
from aws_lambda_powertools.utilities import parameters

dynamodb_provider = parameters.DynamoDBProvider(
    table_name="my-table",
    key_attr="MyKeyAttr",
    sort_attr="MySortAttr",
    value_attr="MyvalueAttr"
)

def handler(event, context):
    value = dynamodb_provider.get("my-parameter")

AppConfigProvider

Using AppConfigProvider
1
2
3
4
5
6
7
8
9
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

config = Config(region_name="us-west-1")
appconf_provider = parameters.AppConfigProvider(environment="my_env", application="my_app", config=config)

def handler(event, context):
    # Retrieve a single secret
    value: bytes = appconf_provider.get("my_conf")

Create your own provider

You can create your own custom parameter store provider by inheriting the BaseProvider class, and implementing both _get() and _get_multiple() methods to retrieve a single, or multiple parameters from your custom store.

All transformation and caching logic is handled by the get() and get_multiple() methods from the base provider class.

Here is an example implementation using S3 as a custom parameter store:

Creating a S3 Provider to fetch parameters
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
import copy

from aws_lambda_powertools.utilities import BaseProvider
import boto3

class S3Provider(BaseProvider):
    bucket_name = None
    client = None

    def __init__(self, bucket_name: str):
        # Initialize the client to your custom parameter store
        # E.g.:

        self.bucket_name = bucket_name
        self.client = boto3.client("s3")

    def _get(self, name: str, **sdk_options) -> str:
        # Retrieve a single value
        # E.g.:

        sdk_options["Bucket"] = self.bucket_name
        sdk_options["Key"] = name

        response = self.client.get_object(**sdk_options)
        return

    def _get_multiple(self, path: str, **sdk_options) -> Dict[str, str]:
        # Retrieve multiple values
        # E.g.:

        list_sdk_options = copy.deepcopy(sdk_options)

        list_sdk_options["Bucket"] = self.bucket_name
        list_sdk_options["Prefix"] = path

        list_response = self.client.list_objects_v2(**list_sdk_options)

        parameters = {}

        for obj in list_response.get("Contents", []):
            get_sdk_options = copy.deepcopy(sdk_options)

            get_sdk_options["Bucket"] = self.bucket_name
            get_sdk_options["Key"] = obj["Key"]

            get_response = self.client.get_object(**get_sdk_options)

            parameters[obj["Key"]] = get_response["Body"].read().decode()

        return parameters

Deserializing values with transform parameter

For parameters stored in JSON or Base64 format, you can use the transform argument for deserialization.

Info

The transform argument is available across all providers, including the high level functions.

1
2
3
4
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    value_from_json = parameters.get_parameter("/my/json/parameter", transform="json")

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
from aws_lambda_powertools.utilities import parameters

ssm_provider = parameters.SSMProvider()

def handler(event, context):
    # Transform a JSON string
    value_from_json = ssm_provider.get("/my/json/parameter", transform="json")

    # Transform a Base64 encoded string
    value_from_binary = ssm_provider.get("/my/binary/parameter", transform="binary")

Partial transform failures with get_multiple()

If you use transform with get_multiple(), you can have a single malformed parameter value. To prevent failing the entire request, the method will return a None value for the parameters that failed to transform.

You can override this by setting the raise_on_transform_error argument to True. If you do so, a single transform error will raise a TransformParameterError exception.

For example, if you have three parameters, /param/a, /param/b and /param/c, but /param/c is malformed:

Raising TransformParameterError at first malformed parameter
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
from aws_lambda_powertools.utilities import parameters

ssm_provider = parameters.SSMProvider()

def handler(event, context):
    # This will display:
    # /param/a: [some value]
    # /param/b: [some value]
    # /param/c: None
    values = ssm_provider.get_multiple("/param", transform="json")
    for k, v in values.items():
        print(f"{k}: {v}")

    try:
        # This will raise a TransformParameterError exception
        values = ssm_provider.get_multiple("/param", transform="json", raise_on_transform_error=True)
    except parameters.exceptions.TransformParameterError:
        ...

Auto-transform values on suffix

If you use transform with get_multiple(), you might want to retrieve and transform parameters encoded in different formats.

You can do this with a single request by using transform="auto". This will instruct any Parameter to to infer its type based on the suffix and transform it accordingly.

Info

transform="auto" feature is available across all providers, including the high level functions.

Deserializing parameter values based on their suffix
1
2
3
4
5
6
from aws_lambda_powertools.utilities import parameters

ssm_provider = parameters.SSMProvider()

def handler(event, context):
    values = ssm_provider.get_multiple("/param", transform="auto")

For example, if you have two parameters with the following suffixes .json and .binary:

Parameter name Parameter value
/param/a.json [some encoded value]
/param/a.binary [some encoded value]

The return of ssm_provider.get_multiple("/param", transform="auto") call will be a dictionary like:

1
2
3
4
{
    "a.json": [some value],
    "b.binary": [some value]
}

Passing additional SDK arguments

You can use arbitrary keyword arguments to pass it directly to the underlying SDK method.

1
2
3
4
5
6
7
from aws_lambda_powertools.utilities import parameters

secrets_provider = parameters.SecretsProvider()

def handler(event, context):
    # The 'VersionId' argument will be passed to the underlying get_secret_value() call.
    value = secrets_provider.get("my-secret", VersionId="e62ec170-6b01-48c7-94f3-d7497851a8d2")

Here is the mapping between this utility's functions and methods and the underlying SDK:

Provider Function/Method Client name Function name
SSM Parameter Store get_parameter ssm get_parameter
SSM Parameter Store get_parameters ssm get_parameters_by_path
SSM Parameter Store SSMProvider.get ssm get_parameter
SSM Parameter Store SSMProvider.get_multiple ssm get_parameters_by_path
Secrets Manager get_secret secretsmanager get_secret_value
Secrets Manager SecretsManager.get secretsmanager get_secret_value
DynamoDB DynamoDBProvider.get dynamodb (Table resource)
DynamoDB DynamoDBProvider.get_multiple dynamodb (Table resource)
App Config get_app_config appconfig get_configuration

Bring your own boto client

You can use boto3_client parameter via any of the available Provider Classes. Some providers expect a low level boto3 client while others expect a high level boto3 client, here is the mapping for each of them:

Provider Type Boto client construction
SSMProvider low level boto3.client("ssm")
SecretsProvider low level boto3.client("secrets")
AppConfigProvider low level boto3.client("appconfig")
DynamoDBProvider high level boto3.resource("dynamodb")

Bringing them together in a single code snippet would look like this:

Example: passing a custom boto3 client for each provider
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
import boto3
from botocore.config import Config

from aws_lambda_powertools.utilities import parameters

config = Config(region_name="us-west-1")

# construct boto clients with any custom configuration
ssm = boto3.client("ssm", config=config)
secrets = boto3.client("secrets", config=config)
appconfig = boto3.client("appconfig", config=config)
dynamodb = boto3.resource("dynamodb", config=config)

ssm_provider = parameters.SSMProvider(boto3_client=ssm)
secrets_provider = parameters.SecretsProvider(boto3_client=secrets)
appconf_provider = parameters.AppConfigProvider(boto3_client=appconfig, environment="my_env", application="my_app")
dynamodb_provider = parameters.DynamoDBProvider(boto3_client=dynamodb, table_name="my-table")
When is this useful?

Injecting a custom boto3 client can make unit/snapshot testing easier, including SDK customizations.

Customizing boto configuration

The config , boto3_session, and boto3_client parameters enable you to pass in a custom botocore config object , boto3 session, or a boto3 client when constructing any of the built-in provider classes.

Tip

You can use a custom session for retrieving parameters cross-account/region and for snapshot testing.

When using VPC private endpoints, you can pass a custom client altogether. It's also useful for testing when injecting fake instances.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
from aws_lambda_powertools.utilities import parameters
import boto3

boto3_session = boto3.session.Session()
ssm_provider = parameters.SSMProvider(boto3_session=boto3_session)

def handler(event, context):
    # Retrieve a single parameter
    value = ssm_provider.get("/my/parameter")
    ...

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

boto_config = Config()
ssm_provider = parameters.SSMProvider(config=boto_config)

def handler(event, context):
    # Retrieve a single parameter
    value = ssm_provider.get("/my/parameter")
    ...

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
from aws_lambda_powertools.utilities import parameters
import boto3

boto3_client= boto3.client("ssm")
ssm_provider = parameters.SSMProvider(boto3_client=boto3_client)

def handler(event, context):
    # Retrieve a single parameter
    value = ssm_provider.get("/my/parameter")
    ...

Testing your code

Mocking parameter values

For unit testing your applications, you can mock the calls to the parameters utility to avoid calling AWS APIs. This can be achieved in a number of ways - in this example, we use the pytest monkeypatch fixture to patch the parameters.get_parameter method:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
from src import index

def test_handler(monkeypatch):

    def mockreturn(name):
        return "mock_value"

    monkeypatch.setattr(index.parameters, "get_parameter", mockreturn)
    return_val = index.handler({}, {})
    assert return_val.get('message') == 'mock_value'

1
2
3
4
5
6
from aws_lambda_powertools.utilities import parameters

def handler(event, context):
    # Retrieve a single parameter
    value = parameters.get_parameter("my-parameter-name")
    return {"message": value}

If we need to use this pattern across multiple tests, we can avoid repetition by refactoring to use our own pytest fixture:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
import pytest

from src import index

@pytest.fixture
def mock_parameter_response(monkeypatch):
    def mockreturn(name):
        return "mock_value"

    monkeypatch.setattr(index.parameters, "get_parameter", mockreturn)

# Pass our fixture as an argument to all tests where we want to mock the get_parameter response
def test_handler(mock_parameter_response):
    return_val = index.handler({}, {})
    assert return_val.get('message') == 'mock_value'

Alternatively, if we need more fully featured mocking (for example checking the arguments passed to get_parameter), we can use unittest.mock from the python stdlib instead of pytest's monkeypatch fixture. In this example, we use the patch decorator to replace the aws_lambda_powertools.utilities.parameters.get_parameter function with a MagicMock object named get_parameter_mock.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
from unittest.mock import patch
from src import index

# Replaces "aws_lambda_powertools.utilities.parameters.get_parameter" with a Mock object
@patch("aws_lambda_powertools.utilities.parameters.get_parameter")
def test_handler(get_parameter_mock):
    get_parameter_mock.return_value = 'mock_value'

    return_val = index.handler({}, {})
    get_parameter_mock.assert_called_with("my-parameter-name")
    assert return_val.get('message') == 'mock_value'

Clearing cache

Parameters utility caches all parameter values for performance and cost reasons. However, this can have unintended interference in tests using the same parameter name.

Within your tests, you can use clear_cache method available in every provider. When using multiple providers or higher level functions like get_parameter, use clear_caches standalone function to clear cache globally.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
import pytest

from src import app


@pytest.fixture(scope="function", autouse=True)
def clear_parameters_cache():
    yield
    app.ssm_provider.clear_cache() # This will clear SSMProvider cache

@pytest.fixture
def mock_parameter_response(monkeypatch):
    def mockreturn(name):
        return "mock_value"

    monkeypatch.setattr(app.ssm_provider, "get", mockreturn)

# Pass our fixture as an argument to all tests where we want to mock the get_parameter response
def test_handler(mock_parameter_response):
    return_val = app.handler({}, {})
    assert return_val.get('message') == 'mock_value'

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
import pytest

from aws_lambda_powertools.utilities import parameters
from src import app


@pytest.fixture(scope="function", autouse=True)
def clear_parameters_cache():
    yield
    parameters.clear_caches() # This will clear all providers cache

@pytest.fixture
def mock_parameter_response(monkeypatch):
    def mockreturn(name):
        return "mock_value"

    monkeypatch.setattr(app.ssm_provider, "get", mockreturn)

# Pass our fixture as an argument to all tests where we want to mock the get_parameter response
def test_handler(mock_parameter_response):
    return_val = app.handler({}, {})
    assert return_val.get('message') == 'mock_value'

1
2
3
4
5
6
7
8
9
from aws_lambda_powertools.utilities import parameters
from botocore.config import Config

ssm_provider = parameters.SSMProvider(config=Config(region_name="us-west-1"))


def handler(event, context):
    value = ssm_provider.get("/my/parameter")
    return {"message": value}
Last update: 2022-05-19
Back to top