Env not propagated to cronjobs (was: Usage metrics not working)

[ktuite]

QA reported that the most recent usage metrics aren't working

[brontolosone]

The prime primitive to convey configuration to the entrypoint of the Docker container, is the env. Cron is written for a different scenario: it is to be run from init as a daemon, and (especially since it runs jobs for multiple users) should not propagate its environment to its jobs.
Resulting in the situation that the DB env vars configuration setup that went in with #1671, are not available to the processes running as cronjobs, and that seems to be the root of the problem.

As discussed on the team meet a few hours ago: we want some variant of solution 1b from the slack thread, which involves making the DB env available to these jobs (just the DB env for now, not the other env vars at this point, even though they probably should be included in the future so that the execution environments are in sync).