Boto can be configured in multiple ways. Regardless of the source or sources that you choose, you musthave AWS credentials and a region set in order to make requests.
There are two types of configuration data in boto3: credentials and non-credentials. Credentials include items such as aws_access_key_id, aws_secret_access_key, and aws_session_token. Non-credential configuration includes items such as which region to use or which addressing style to use for Amazon S3. The distinction between credentials and non-credentials configuration is important because the lookup process is slightly different. Boto3 will look in several additional locations when searching for credentials that do not apply when searching for non-credential configuration.
The mechanism in which boto3 looks for credentials is to search through a list of possible locations and stop as soon as it finds credentials. The order in which Boto3 searches for credentials is:
- Passing credentials as parameters in the boto.client() method
- Passing credentials as parameters when creating a Session object
- Environment variables
- Shared credential file (~/.aws/credentials)
- AWS config file (~/.aws/config)
- Assume Role provider
- Boto2 config file (/etc/boto.cfg and ~/.boto)
- Instance metadata service on an Amazon EC2 instance that has an IAM role configured.
Each of those locations is discussed in more detail below.
The first option for providing credentials to boto3 is passing them as parameters when creating clients or when creating a Session. For example:
client = boto3.client(
# Or via the Session
session = boto3.Session(
where ACCESS_KEY, SECRET_KEY and SESSION_TOKEN are variables that contain your access key, secret key, and optional session token. Note that the examples above do not have hard coded credentials. We do not recommend hard coding credentials in your source code. For example:
# Do not hard code credentials
client = boto3.client(
# Hard coded strings as credentials, not recommended.
Valid uses cases for providing credentials to the client() method and Session objects include:
- Retrieving temporary credentials using AWS STS (such as sts.get_session_token()).
- Loading credentials from some external location, e.g the OS keychain.
Boto3 will check these environment variables for credentials:
- The access key for your AWS account.
- The secret key for your AWS account.
- The session key for your AWS account. This is only needed when you are using temporary credentials. The AWS_SECURITY_TOKEN environment variable can also be used, but is only supported for backwards compatibility purposes. AWS_SESSION_TOKEN is supported by multiple AWS SDKs besides python.
AWS Config File
Boto3 can also load credentials from ~/.aws/config. You can change this default location by setting the AWS_CONFIG_FILE environment variable. The config file is an INI format, with the same keys supported by the shared credentials file. The only difference is that profile sections must have the format of [profileprofile-name], except for the default profile. For example:
# Example ~/.aws/config file.
The reason that section names must start with profile in the ~/.aws/config file is because there are other sections in this file that are permitted that aren’t profile configurations.
Assume Role Provider
This is a different set of credentials configuration than using IAM roles for EC2 instances, which is discussed in a section below.
Within the ~/.aws/config file, you can also configure a profile to indicate that boto3 should assume a role. When you do this, boto3 will automatically make the corresponding AssumeRole calls to AWS STS on your behalf. It will handle in memory caching as well as refreshing credentials as needed.
You can specify the following configuration values for configuring an IAM role in boto3:
- role_arn – The ARN of the role you want to assume.
- source_profile – The boto3 profile that contains credentials we should use for the initial AssumeRolecall.
- external_id – A unique identifier that is used by third parties to assume a role in their customers’ accounts. This maps to the ExternalId parameter in the AssumeRole operation. This is an optional parameter.
- mfa_serial – The identification number of the MFA device to use when assuming a role. This is an optional parameter. Specify this value if the trust policy of the role being assumed includes a condition that requires MFA authentication. The value is either the serial number for a hardware device (such as GAHT12345678) or an Amazon Resource Name (ARN) for a virtual device (such as arn:aws:iam::123456789012:mfa/user).
- role_session_name – The name applied to this assume-role session. This value affects the assumed role user ARN (such as arn:aws:sts::123456789012:assumed-role/role_name/role_session_name). This maps to the RoleSessionName parameter in the AssumeRole operation. This is an optional parameter. If you do not provide this value, a session name will be automatically generated.
If you do not have MFA authentication required, then you only need to specify a role_arn and a source_profile.
When you specify a profile that has IAM role configuration, boto3 will make an AssumeRole call to retrieve temporary credentials. Subsequent boto3 API calls will use the cached temporary credentials until they expire, in which case boto3 will automatically refresh credentials. boto3 does not write these temporary credentials to disk. This means that temporary credentials from the AssumeRole calls are only cached in memory within a single Session. All clients created from that session will share the same temporary credentials.
If you specify an mfa_serial, then the first time an AssumeRole call is made, you will be prompted to enter the MFA code. Your code will block until you enter your MFA code. You’ll need to keep this in mind if you have an mfa_serial configured but would like to use boto3 in some automated script.
Below is an example configuration for the minimal amount of configuration needed to configure an assume role profile:
# In ~/.aws/credentials:
# In ~/.aws/config
See Using IAM Roles for general information on IAM roles.
Boto3 will attempt to load credentials from the Boto2 config file. It first checks the file pointed to by BOTO_CONFIG if set, otherwise it will check /etc/boto.cfg and ~/.boto. Note that only the [Credentials]section of the boto config file is used. All other configuration data in the boto config file is ignored. Example:
# Example ~/.boto file
aws_access_key_id = foo
aws_secret_access_key = bar
This credential provider is primarily for backwards compatibility purposes with boto2.
If you are running on Amazon EC2 and no credentials have been found by any of the providers above, boto3 will try to load credentials from the instance metadata service. In order to take advantage of this feature, you must have specified an IAM role to use when you launched your EC2 instance. For more information on how to configure IAM roles on EC2 instances, see the IAM Roles for Amazon EC2 guide.
Note that if you’ve launched an EC2 instance with an IAM role configured, there’s no explicit configuration you need to set in boto3 to use these credentials. Boto3 will automatically use IAM role credentials if it does not find credentials in any of the other places listed above.
Best Practices for Configuring Credentials
If you’re running on an EC2 instance, use AWS IAM roles. See the IAM Roles for Amazon EC2 guide for more information on how to set this up.