serverless-config 0.3.1

A configuration client for AWS serverless Python systems

A simple configuration client for AWS serverless Python systems.

There is also a jvm version.

Installation

pip install serverless-config

or

pipenv install serverless-config

AWS lambda includes boto3 in its environment, so serverless-config does not include it as a dependency in order to decrease the deployment package size. If you wish to use serverless-config locally, be sure to install boto3 as well.

Quickstart

from serverless_config import default_config
config = default_config()

string_prop = config.get_str('string_prop')
int_prop = config.get_int('missing_int_prop', default_value=123)
secret_prop = config.get_str('secret_prop', WithDecryption=True)

The default config will search for a parameter with the following order of precedence: System Environment, AWS SSM Parameter Store. You can learn more about them below.

Supported Config Sources

System Environment

The System environment is a good place to store microservice-specific parameters. They are set on the lambda function itself.

from serverless_config import EnvConfig
config = EnvConfig()
config.get_str('string_prop')

AWS SSM Parameter Store

SSM is perfect for storing parameters that are shared across microservices, and for storing encrypted secrets. It is fully managed, and does not require any configuration to get started.

Note: the IAM role requires the AmazonSSMReadOnlyAccess policy to get properties from SSM.

from serverless_config import SsmConfig
config = SsmConfig()
config.get_str('string_prop')

A secret can optionally be decrypted in transit. That way, you do not need to worry about configuring your IAM role for access to the KMS Key.

from serverless_config import SsmConfig
config = SsmConfig()
config.get_str('string_prop', WithDecryption=True)

Composite Configs

The default_config will first search in the system environment. If the parameter is not there, then it will search in AWS SSM.

from serverless_config import default_config
config = default_config()

Custom Configs

You can even implement your own custom configs and composite configs!

from serverless_config import ConfigBase, CompositeConfig, EnvConfig

class DictConfig(ConfigBase):

    def __init__(self, prop_dict):
        self.prop_dict = prop_dict

    def get_str(prop_name, default_value=None):
        if prop_name in self.prop_dict:
            return self.prop_dict[prop_name]
        elif default_value:
            return default_value

        # You must raise a ValueError if the property is not found
        raise ValueError('Property not found: ' + prop_name)

# You can make a standalone custom config
props = dict(foo='bar', toll='troll')
map_config = DictConfig(props)

# And you can make a custom composite config with your new config
custom_config = CompositeConfig(map_config, EnvConfig())

Caching

The default_config will cache properties for 5 minutes. If you wish to use a specific or custom config, you can wrap the CachedConfig around it.

from serverless_config import default_config

config = default_config()
value = config.get_str('prop')  # getting value from env and ssm
value = config.get_str('prop')  # getting cached value

from datetime import timedelta
from serverless_config import EnvConfig, SsmConfig, CachedConfig, CompositeConfig

CachedConfig(SsmConfig())  # config with default 5 minute cache duration
CachedConfig(SsmConfig(), timedelta(hours=1)  # config with 1 hour cache duration
CachedConfig(CompositeConfig(SsmConfig(), EnvConfig()))  # you can even cache a composite config!