Use rawConnection or Pool in node-redshift for lambda?

[Atik1608]

I am stuck at a scenario in which whether to use rawConnection or deafult connection-pooling of node-redshift. The application is a Lambda function which is going to insert the data into redshift table and it's going to be at scale. I want to make sure that I use optimized way of connection and follow best practices! Any help would be appreciated.