www: add initial cloudflare log-collect tooling

[rvagg]

Includes the two commits in #976 which should land shortly.

As I said there:

Adds the log-collect stuff that's on there collecting logs for us from cloudflare and doing some basic filtering. There's also some stuff in there to convert the json format to nginx format but it's not being used for anything currently. The log-collect work is only partially complete since we're collecting but not using it for metrics generation yet. We still need to (a) confirm that we're successfully collecting everything with this procedure (logshare isn't the most robust tool unfortunately, or maybe it's their API, hence the sanity-check script in here), and (b) either do the json->nginx conversion for metrics, or make metrics scripts able to read the json files as they are. Either way, this PR reflects current state so is good to merge IMO.

I haven't tested this .. a bit of a stab in the dark, I'm not willing to run this against the live server.

I'm not sure that I'm going to be the person to finish this work but there's only a small number of people we can give access to the webserver. So what I'm going to have to do is spin up a separate server that has all of this log collecting stuff, metrics, old logs and some sensitive stuff (cloudflare key for one) and give someone, or someones access to it. I believe @refack volunteered to have a go at it when I raised this previously somewhere.

[rvagg]

forgot about #679 which is essentially the same. I've closed that but there's some notes there that are relevant to this

[rvagg]

see #2025