When you jumped between these AWS instance types did you tune / monitor to see what resources were becoming constrained? Or can you provide some details on netw

When you jumped between these AWS instance types did you tune / monitor to see what resources were becoming constrained? Or can you provide some details on network / VPC setup and changes to http connection pools in the AWS SDK? Since timescaledb is communicating over HTTP through the AWS client, I'm wondering how much you can tune to eek out performance there. I've been burned so many times across AWS services on classes of error like this that I'm just casually curious, not that anything was done "wrong".