The discussion naturally turned out to discuss concurrency, Docker, Kubernetes, and the unreliability of the underlying network

The most interesting so far IMO has been about Postgres Partitioning Best Practices. Notable mention for pg_partman which I didn’t know about

Alternatively, you use the latter methods and parse the infamous {"$serde_json::private::Number": "value"} map manually, which is not great either. Both the map key identifier and utils to work with it are not part of the library public API

See github.com/serde-rs/jso...

Okay, so apparently there is a workaround: you deserialize your JSON string into serde_json::Value first, and then you convert that into the struct of your choice by using serde_json::from_value.

Obviously it is not as efficient as using serde_json:from_slice or serde_json::from_str