chore: move committeeclash pg to host pg

[SilasPeters]

Yay

[SilasPeters]

This is intrusive though

[SilasPeters]

@rensvmoorsel once this PR is merged, we will require copying over all data from the containerisd postgress instance to the public postgres instance. A pg_dump could suffice, but we have to make sure not to override anything already existing in production. We could use staging as a test.

The database into which the data needs to be copied is best made by deplying the changes in this PR, meaning that there is a brief moment where committeeclash must be made unavailable through nginx (maintenance mode) because we then are manually filling the used production database which is then empty.

[KoenDR06]

TODO: fix auth issue

@SilasPeters is this TODO resolved?

[SilasPeters]

@KoenDR06 yes

[KoenDR06]

While deploying the committeeclash role just now, the server seemed to not respond to any requests, I even got a 500 from koala. If this was an issue before this PR it's good to be merged in my opinion.

[SilasPeters]

I experienced the same earlier. Not sure what causes it, I suspect a shortage of memory

[KoenDR06]

Seems like it was already an issue, LGTM!