update whodb docs

[modelorona]

Hello! It's me again.

I updated these to have more information about all the available env variables and how to configure profiles + ai providers.

Realised it might be too much information but let me know! Have got this on our docs as well but need to rebuild them so it's not live - when that is, can link to it directly and clean this up.

Let me know what works best for you guys!

[Joulinar]

Hi, thanks for the PR. That might indeed be a bit too much. Perhaps we should limit ourselves to the core points. Maybe 3-5 at most. For the rest, we could refer to your official documentation? That would save us from constantly updating two overview pages.

[modelorona]

No worries, I'll update when I rebuild docs.

I can link to the example of configuring connections + ai providers. For the env variables, do you have a preference on which core ones to keep?

[Joulinar]

I was thinking about

PORT, WHODB_LOG_LEVEL, WHODB_LOG_FORMAT, WHODB_LOG_FILE and WHODB_ACCESS_LOG_FILE

But that's just my personal opinion. Let's see what the others have to say about it.

[MichaIng]

Also WHODB_LOG_FORMAT isn't that relevant in our case, since the process is started as systemd service, so all output goes to journal/syslog. That is, unless WHODB_LOG_FILE/WHODB_ACCESS_LOG_FILE is used, where it applies as well, right?

Regarding WHODB_LOG_FILE: it probably makes sense to clarify that the default (unset) means logs go to STDOUT (=>journal in case of systemd) and the value default enables file logging with a default path, but is not actually the default value of this setting? It is probably a bit confusing 😄.

Related: it probably makes sense to somehow differentiate between default values, and initial values used in our whodb.env. E.g WHODB_LOG_LEVEL default is info, while only in our case it is set to warning in our initial whodb.env. Similar with the port IIRC.

And I agree with @Joulinar that we should not have a copy of full environment variable docs here, but keep a subset of most relevant ones. Access port and restrictions, logging, pre-connecting to databases is all something I would immediately want to look for, while e.g. AI provider settings is IMO not relevant to get started. For this, a link to whichever upstream docs/wiki page or related markdown document makes sense. So the maintenance efforts our end is reduced.