Fix static parser tracking logic

[jtcohen6]

Background

I noticed a few weird things this morning:

• We have way too many cannot_parse results in anonymous usage tracking for v1 — many more than I'd expect if we're only sampling one out of every 50k models
• Using latest v1, dbt --no-partial-parse parse ("cold parse") in our internal analytics project is taking 15-16 seconds, instead of 7-8 seconds. Indeed, that extra time is spent in tracking:

Fix

We're firing a tracking event every single time the stable parser cannot_parse (or has_banned_macro). Instead, we should only populate the tracking result if we're sampling. When I make this change, a "cold parse" of our internal analytics project takes 7-8 seconds again. Hooray!

While here:

• Consolidated the logic, so this check + result population happens as part of the "fall back to jinja" conditional branch
• We should be checking experimental_sample, too, and firing a differently coded event depending (0xvs. 8x)

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

[gshank]

Looks good to me.

[nathaniel-may]

Oof. good catch! thanks for fixing this!