[FEA] Extract and report Spark Connect session and operation metadata

[sayedbilalbari]

Parent issue: #2058

Summary

Implement full Spark Connect metadata extraction and reporting on top of the parser awareness added in earlier phases.

This is the Phase 3 work item from #2058.

Problem

Once Connect events are accepted, the tools still need to interpret them to provide user/session attribution, operation lifecycle timing, and operation-level failure metadata. This is the feature-complete Spark Connect support layer.

Scope

• Parse Connect session lifecycle events for sessionId and userId
• Parse Connect operation lifecycle events for operationId, jobTag, timing markers, row counts, and failure metadata
• Correlate operation events with SQL executions/jobs via jobTag
• Support per-session config tracking by combining session correlation with modifiedConfigs
• Surface relevant metadata in qualification/profiling outputs where appropriate
• Consider multi-user reporting or attribution breakdowns where the data model supports it

Acceptance Criteria

• Session lifecycle metadata is captured and queryable
• Operation lifecycle timing can be reconstructed from Connect events
• Operation failures/cancellations are attributable in tool output or internal models
• Connect operations can be correlated to SQL executions/jobs through jobTag
• Tests cover session attribution, timing reconstruction, and failure handling

Notes

Relevant analysis in repo:

• core/docs/spark-connect-events-analysis.md
• core/docs/spark-connect-operation-started-examples.md