Hapi.js, ZSTD and the Roadblocks

Regarding Zstd support, I don't think it makes sense to make an exception for the policy around node API feature stability. Especially because there is a standard way to integrate custom compressors.

For brotli, I still don't think we need to include it in core, but I'm not opposed to it, if done properly. This implementation still has several issues, most pressingly a default compression level of 11, which is way too high for just-in-time compression. The default should probably not be more computationally expensive than the default gzip option. A quick test on a 300KB json file shows gzip to take 20 ms, while brotli takes 335 ms! For this file, brotli needs to use level 5 to have comparable speed. I would probably choose 4 as default, improving both compression speed and a size.

I wish you would make a separate PR for the priority option, which provides a meaningful feature by itself.