When to use boto3 sessions explicitly

Question

By default boto3 creates sessions whenever required, according to the documentation

it is possible and recommended to maintain your own session(s) in some scenarios

My understanding is if I use a session created by me I can reuse the same session across the application instead of boto3 automatically creating multiple sessions or if I want to pass credentials from code.

Has anyone ever maintained sessions on their own? If yes what was the advantage that it provided apart from the one mentioned above.

secrets_manager = boto3.client('secretsmanager')

session = boto3.session.Session()
secrets_manager = session.client('secretsmanager')

Is there any advantage of using one over the other and which one is recommended in this case.

References: https://boto3.amazonaws.com/v1/documentation/api/latest/guide/session.html

score 2 · Answer 1 · answered Jan 01 '20 at 08:05

I have seen the second method used when you wish to provide specific credentials without using the standard Credentials Provider Chain.

For example, when assuming a role, you can use the new temporary to create a session, then create a client from the session.

From boto3 sessions and aws_session_token management:

import boto3

role_info = {
    'RoleArn': 'arn:aws:iam::<AWS_ACCOUNT_NUMBER>:role/<AWS_ROLE_NAME>',
    'RoleSessionName': '<SOME_SESSION_NAME>'
}

client = boto3.client('sts')
credentials = client.assume_role(**role_info)

session = boto3.session.Session(
    aws_access_key_id=credentials['Credentials']['AccessKeyId'],
    aws_secret_access_key=credentials['Credentials']['SecretAccessKey'],
    aws_session_token=credentials['Credentials']['SessionToken']
)

You could then use: s3 = session.client('s3')

score 0 · Answer 2 · answered Oct 12 '21 at 02:59

Here is an example where I needed to use the session object with both Boto3 and AWS Datawrangler to set the region for both:

REGION = os.environ.get("REGION")

session = boto3.Session(region_name=REGION)
client_rds = session.client("rds-data")

df = wr.s3.read_parquet(path=path, boto3_session=session)

When to use boto3 sessions explicitly

2 Answers2