jetmonv2 adapter: classify truncated-body failures as partial_response

[heydemoura]

Summary

• Map Jetmon v2 error_code=8 (truncated body / partial payload) to partial_response in the adapter normalization path.
• Treat partial_response as a first-class normalized outage type (not raw-classification reporting).
• Keep source + docs aligned by updating adapter mapping/tests and event taxonomy docs.

Validation

• go test ./internal/adapter/jetmonv2/... -> pass (ok, cached)
• go test ./internal/adapter/... -> pass (all adapter packages ok; base internal/adapter package has no test files)

End-to-End Evidence

• Run ID 37e518b56abe97c8b9daad1025ebe88e shows monitor_reports carrying raw and normalized partial_response for the truncated-body scenario.
• This confirms raw ingestion + normalized classification are both emitted as expected.

[chrisbliss18]

Thanks for this. The uptime-bench side looks directionally right: mapping Jetmon v2 error_code=8 to partial_response makes sense, and the added test covers the important case where Jetmon reports Seems Down with http_code=200 but a body-integrity error.

I'd like to hold this until Automattic/jetmon#87 settles the Jetmon-side of things. This PR currently depends on error_code=8 being the stable external signal for truncated/partial response failures. If the Jetmon PR changes that shape while addressing the review feedback there, for example by renaming the internal class, changing the numeric code, or exposing a different metadata field, this adapter mapping would need to follow it.

In the meantime, one small cleanup: docs/adapters.md still lists the normalized vocabulary without partial_response, even though adapter.go and docs/events.md now include it. Please add partial_response there too so the adapter-author guidance stays consistent.

Once the Jetmon metadata contract is confirmed, I think this should be straightforward to land.