Wrapping the asynchronous Meilisearch API to run on AWS Lambda with EFS storage.
Meilisearch is mainly asynchronous, and is designed to run on a server that is always on.
To experiment with cost optimizations, we would like to run it on a Lambda function, which is ephemeral and may be shut down at any time, backed by EFS for perpetual storage of the graph data itself.
This project provides a wrapper around Meilisearch that allows it to run on an EFS-linked Lambda function, by wrapping the asynchronous calls in a wrapper that waits for the operations to complete before returning.
Find an example Terraform setup for AWS Lambda + EFS in the docs/ folder.
It works on my machine 🤷♂️. Not entirely sure about multiple concurrent requests. This is a proof of concept. Contributions are welcome!