Hey! I'm wondering what would be the most efficient way to upload around 80M documents into a collection? I'm using the /import endpoint but I'm not sure if sending multiple requests at once is faster or more data at once. Our server is 32gb of ram with 4 cores. Right now I'm fetching 500k rows from mysql converting to jsonl and posting. Right now it takes about 15s per batch using go. I set the batch_size to 500k as well.