tracking: flow incremental read

[discord9]

Summary

Support incremental read in flow engine — when a flow runs, only process the new data since the last run, rather than reprocessing the entire table. This significantly reduces recomputation overhead for continuous flow queries.

Related PRs

• feat: add incremental read context and scan boundaries #7848 — feat: add incremental read context and scan boundaries
• feat(flow): inc query scan bind seq #7879 — feat(flow): inc query scan bind seq
• feat: remote dyn filter basics #7979 — feat: remote dyn filter basics
• feat: add seq watermark in record batch metrics #8015 — feat: add seq watermark in record batch metrics
• feat: flow inc query terminal metrics transport #8045 — feat: flow inc query terminal metrics transport
• feat: inc query join rewrite helper #8108 — feat: inc query join rewrite helper (reviewing)
• flownode-side incremental read support (not yet submitted)

Component Breakdown

Component	Description	Status
Scan boundaries / context	Defines IncrementalReadContext and scan range boundaries so the flow knows what data to scan	✅
Scan bind sequence	Binds scan sequence numbers to flow queries for incremental processing	✅
Remote dyn filter	Passes dynamic filter expressions to remote scans	✅
Seq watermark metrics	Records sequence watermark in record batch metrics for progress tracking	✅
Terminal metrics transport	Transports flow incremental query terminal metrics	✅
Join rewrite helper	Rewrites aggregate delta queries to LEFT JOIN with sink table for incremental merge	🔄
Flownode integration	Integrate incremental read support into the flownode execution path	🔜