Hey everyone! I'm currently working on a RAG (Retrieval-Augmented Generation) app that utilizes external embeddings and LLM APIs. The code's become too complex for AWS Lambda, so I decided to containerize it and will be running it on Fargate instead. I've already integrated the vector database logic inside the container. Now, I'm looking for advice on the best and most cost-effective way to store the embeddings without resorting to RDS or DynamoDB. EFS has crossed my mind, but I'm wondering if there might be a faster or cheaper alternative. Also, can EFS actually store the container embedding documents, or is it limited to being just a file system?
4 Answers
Don't underestimate RDS! Depending on your needs, it can be very reasonable. If not, consider exploring other storage options outside the default AWS suggestions. You might find a good fit!
Have you thought about asking AI for help? I mean, throwing this question at ChatGPT could yield some interesting ideas or directions you might not have considered yet!
It sounds like you're dealing with some heavy logic if Lambda isn't cutting it. Have you considered using Pinecone? They've got some decent free tier options for vector storage. It might simplify things for you!
Thanks for the suggestion! I'll check it out.
Yeah, I agree with CodeWizard42. Pinecone can be pretty handy and might be a good fit depending on your specific requirements.
Just to clarify, are you looking to avoid using a database entirely or just AWS-managed solutions? Sometimes, managed databases can be more cost-effective than expected depending on usage. Take a closer look at your needs before ruling them out completely!
I'm aiming for the best cost-efficiency possible! I initially thought RDS was my best bet, but it seems pricey.
I’ve been discussing this with ChatGPT all day—this is kind of my last resort!