I am using boto3 to read s3 objects
s3_client = boto3.client('s3', region_name='us-east-1')
obj = s3_client.get_object(Bucket=S3_BUCKET, Key=key)
I am running this via 50-100 threads to access different objects and getting warning :
urllib3.connectionpool - WARNING - Connection pool is full, discarding connection: s3.amazonaws.com
How can I increase the connection pool size?
Is there any better way to access different S3 objects with multiple threads?
Adding max_pool_connections
(default 10) solved it.
client_config = botocore.config.Config(
max_pool_connections=50
)
s3_client = boto3.client('s3', region_name='us-east-1', config=client_config)