I'm working on a configuration for a reverse proxy that's only purpose is to load balance requests between two upstream servers, primarily for redundancy in case one goes down.
I'm curious about the affect of
proxy_buffering, described here.
I'm load balancing requests for large video files, where the largest and most commonly requested "chunk" of a video is around 2.5 Megabytes. Our CDN makes requests for these videos using byte-rages. The video files are much larger than that.
The default for
proxy_buffering is that it's enabled. My questions are:
- What is the benefit for nginx buffering the response from the proxied server? It must have some value if (a) it's an option and (b) it's turned on by default.
- Does it make sense to disable it in my case? Where the requested data is typically quite large?