reader_executor_block_size
Buffer chunk size for the experimentalReaderExecutor: source reads fill nodes of at most this size. Must be between 128 KiB and 40 MiB, the band shared by every ReaderExecutor size.
reader_executor_max_tail_for_drain
Drain bound for the experimentalReaderExecutor: a long source connection dropped within this many bytes of its right bound is read out to the bound first, so it completes and returns to the connection pool reusable instead of counting as an incomplete connection.
reader_executor_min_bytes_for_seek
Forward-gap bound for the experimentalReaderExecutor: a gap up to this is skipped on the open source connection (bridged / read through) instead of issuing a separate source read or reopening. Set near the bandwidth/request cost breakeven so bridging stays cost-positive.
reader_executor_plan_look_ahead
How far ahead the experimentalReaderExecutor resolves cache residency into its held read plan (a cheap probe, not a read), so one resolve serves many windows. The plan pins every cache cell it resolved until the cursor passes it, so this span drives how much memory and how much unevictable cache each concurrent reader holds. It is not an exact cap: a pin covers a whole cache segment or block, so a cell straddling either end of the range is held entire, adding up to one of them per cache tier at each end. Must be between 128 KiB and 40 MiB, the band shared by every ReaderExecutor size, and at least reader_executor_block_size.
reader_executor_use_long_connections
Reuse a bounded long source connection across windows in the experimentalReaderExecutor. A long connection is one whose range exceeds the current read window; when disabled, the executor takes no connection-pool budget and every window opens a short-lived one-shot connection (the stateless path).
reader_executor_window_size
Bytes served per read window by the experimentalReaderExecutor (the unit a read returns). Must be between 128 KiB and 40 MiB, the band shared by every ReaderExecutor size.