Achieve parity with human-readable end-of-test summary

[oleiade]

Summary

Following the definition of the MVP schema for the machine-readable summary, we aim to enrich the summary with the remaining critical data pieces. The goal is to achieve data parity with the standard, human-readable end-of-test summary, making the new format a complete and reliable data source for automation and analysis.

Problem

The Minimum Viable Schema defined in #4870 provides the essential top-level structure but intentionally leaves the detailed contents of thresholds, groups, and scenarios out of the picture. For the new machine-readable summary to be a viable replacement for parsing human-readable text, it must contain the same level of granular data so that users can rely on it for CI/CD checks and reporting.

Final Deliverables

Upon completion of all tasks in this epic, the following will be true:

• The formal JSON schema (if any) is fully updated with detailed, non-placeholder definitions.
• The Go data model code is implemented to populate all new structures.
• The generated JSON summary file contains rich data for thresholds, groups, checks, and scenarios. The official k6 documentation is updated to reflect the complete schema.

Open Questions for Discussion

• Implementation Order: Does the proposed order (1. Thresholds, 2. Groups/Checks, 3. Scenarios) seem logical? Thresholds are arguably the most critical for CI, making them a good first candidate.
• Sub-Issues: Should we create separate, linked GitHub issues for each of the three main tasks above? This would help keep discussions and pull requests highly focused.