Currently we're not caching GitHub Pages (or in fact any of our backend sites). Doing so might reduce the I/O our proxy needs to do in the case where pages haven't changed.
Apparently compression is disabled on the upstream side of the connection due to our use of url rewriting, meaning that we're currently communicating with e.g: GitHub Pages in a non-compressed format. Since we do (I believe) send out compressed data to clients I'm guessing this means we're also compressing everything on the fly all the time, which is also probably not great for workload. (Not that this has actually been an issue thus far).
Currently we're not caching GitHub Pages (or in fact any of our backend sites). Doing so might reduce the I/O our proxy needs to do in the case where pages haven't changed.
Apparently compression is disabled on the upstream side of the connection due to our use of url rewriting, meaning that we're currently communicating with e.g: GitHub Pages in a non-compressed format. Since we do (I believe) send out compressed data to clients I'm guessing this means we're also compressing everything on the fly all the time, which is also probably not great for workload. (Not that this has actually been an issue thus far).
As reported by @prophile in Slack (https://studentrobotics.slack.com/archives/C02BXUAK33M/p1684074664913729)