Open
Description
Given that there is now "official" guidance from MongoDB Inc. on using their official async python library for mongodb, i.e. motor, with FastAPI, explore pros/cons of migrating from using the older, more-stable-but-not-async pymongo library in order to take full performance advantage of FastAPI's async support.
The implication of this would be that a single-container deployment of the runtime would:
- support a higher rate of requests (avoid having to run multiple containers)
- in turn, potentially help avoid having to consider race conditions wrt multiple runtime containers accessing the same backing mongodb instance
Goal: take some time to compare both approaches to better inform a path forward