Report machine uptime as part of the status json output

Currently we only have the `uptime_seconds` field in the `process` information (which will be obviously resetted once the process gets restarted e.g. for an update) I think it would be very useful to also expose the `uptime_seconds` on the `machine` level to see how long the machines are up and running. If I understand all code pieces correctly we should already have everything to expose this information which would probably require some small changes.

Start by tracing the status JSON generation and the existing process-level uptime_seconds field, then identify how machine-level information is assembled. Done means the status JSON also exposes machine.uptime_seconds without losing the existing process value.

Written by the indexing model from the issue text.