Add invocation_id to BigQuery jobs

[mescanne]

This resolves #2808.

Description

It adds the invocation_id into every label for a BigQuery job. The reason this is preferrable is because the labels have a 128 byte limit, and using the query comment could easily breach this limit.

Checklist

• I have signed the CLA
• I have run this code in development and it appears to resolve the stated issue
• This PR includes tests, or tests are not required/relevant for this PR
• I have updated the CHANGELOG.md and added information about my change to the "dbt next" section.

[jtcohen6]

@mescanne Thanks for the contribution, and sorry for my initial hesitations here. The change is really straightforward and, as I think more about it, the addition is uncontroversial.

I don't think this PR fully resolves #2483, which asks for the ability to configure job tags/labels. It definitely resolves the subset functionality contained in #2808, which I've reopened. We may end up deciding that, by using the invocation_id as foreign key between dbt artifacts (run_results.json) and BigQuery query history, we have access to the metadata we need and are no longer in need of more configurable tags/labels.

So my only ask would be that you update the changelog to reflect resolution of #2808 rather than #2483.

[mescanne]

Done -- reflected to be #2808.

The team I'm working with already records DBT runs using the invocation_id into an audit table in BigQuery. So linking it to the INFORMATION_SCHEMA and using this label is a very natural and easy solution.

Another approach I am thinking about is the INFORMATION_SCHEMA jobs table has the raw query as well. You could use BigQuery regex parsing and JSON parsing to extract information from the query_comment. However this is potentially costly (reading all of the queries - potentially big queries - back across all time), and also quite tricky from a SQL perspective. An audit table with your logical attributes for the run combined with joining on invocation_id is fairly straight forward.

Adding in the stage and other step-level information as labels could be good, but one step at a time..

Thanks a lot.