Stream biz_state metric persist to cut peak memory on large BGP tables.

Write JSONL iteratively, flush via chunked PG execute_values, skip mega raw in DB, and yield bgp_route rows instead of building a full list.

Co-authored-by: Cursor <cursoragent@cursor.com>

netx_api/biz_state/collect_runner.py

@@ -2,9 +2,11 @@
 
 from __future__ import annotations
 
+import json
 import logging
 import re
 import threading
+from collections.abc import Iterable
 from concurrent.futures import ThreadPoolExecutor
 from datetime import datetime
 from typing import Any
@@ -152,16 +154,23 @@ def _resolve_collect_profile(profile_id: str):
     return None
 
 
+_METRIC_CHUNK = 2000
+
+
 def _persist_lldp_rows(
     db,
     *,
     batch: BizStateBatch,
     cmd_row: BizStateBatchCommand,
-    records: list[dict[str, Any]],
+    records: Iterable[dict[str, Any]],
 ) -> int:
+    """Insert LLDP rows from an iterable (chunk-friendly)."""
     n = 0
     seen: set[tuple[str, str, str]] = set()
+    buf: list[BizStateLldpNeighbor] = []
     for rec in records:
+        if not isinstance(rec, dict):
+            continue
         local_if = str(rec.get("local_if") or "").strip()[:128]
         remote_sys = str(rec.get("remote_sys") or "").strip()[:256]
         remote_if = str(rec.get("remote_if") or "").strip()[:128]
@@ -171,7 +180,7 @@ def _persist_lldp_rows(
         if key in seen:
             continue
         seen.add(key)
-        db.add(
+        buf.append(
             BizStateLldpNeighbor(
                 id=uuid4().hex,
                 batch_id=batch.id,
@@ -187,6 +196,190 @@ def _persist_lldp_rows(
             )
         )
         n += 1
+        if len(buf) >= _METRIC_CHUNK:
+            db.add_all(buf)
+            db.flush()
+            buf.clear()
+    if buf:
+        db.add_all(buf)
+        db.flush()
+    return n
+
+
+def _db_is_postgres(db) -> bool:
+    try:
+        bind = db.get_bind() if hasattr(db, "get_bind") else getattr(db, "bind", None)
+        name = str(getattr(getattr(bind, "dialect", None), "name", "") or "").lower()
+        return name in ("postgresql", "postgres")
+    except Exception:
+        return False
+
+
+def _persist_metric_chunk_bulk(
+    db,
+    *,
+    batch: BizStateBatch,
+    cmd_row: BizStateBatchCommand,
+    metric_id: str,
+    chunk: list[dict[str, Any]],
+    seq_start: int,
+) -> int:
+    if not chunk:
+        return 0
+    now = _utcnow()
+    buf = [
+        {
+            "id": uuid4().hex,
+            "batch_id": batch.id,
+            "batch_command_id": cmd_row.id,
+            "task_id": batch.task_id,
+            "ne_id": batch.ne_id,
+            "metric_id": metric_id,
+            "seq": seq_start + i,
+            "data_json": dict(rec),
+            "collected_at": now,
+        }
+        for i, rec in enumerate(chunk)
+        if isinstance(rec, dict) and rec
+    ]
+    if not buf:
+        return 0
+    db.bulk_insert_mappings(BizStateMetricRow, buf)
+    db.flush()
+    return len(buf)
+
+
+def _persist_metric_chunk_copy(
+    db,
+    *,
+    batch: BizStateBatch,
+    cmd_row: BizStateBatchCommand,
+    metric_id: str,
+    chunk: list[dict[str, Any]],
+    seq_start: int,
+) -> int:
+    """Postgres fast path via execute_values; falls back to bulk on error."""
+    if not chunk:
+        return 0
+    now = _utcnow()
+    rows: list[tuple[Any, ...]] = []
+    for i, rec in enumerate(chunk):
+        if not isinstance(rec, dict) or not rec:
+            continue
+        rows.append(
+            (
+                uuid4().hex,
+                batch.id,
+                cmd_row.id,
+                batch.task_id or "",
+                batch.ne_id or "",
+                metric_id,
+                seq_start + i,
+                json.dumps(rec, ensure_ascii=False, default=str, separators=(",", ":")),
+                now,
+            )
+        )
+    if not rows:
+        return 0
+    try:
+        from psycopg2.extras import execute_values  # type: ignore
+    except ImportError:
+        return _persist_metric_chunk_bulk(
+            db,
+            batch=batch,
+            cmd_row=cmd_row,
+            metric_id=metric_id,
+            chunk=chunk,
+            seq_start=seq_start,
+        )
+    # Unwrap SQLAlchemy connection → DBAPI (psycopg2) connection
+    sa_conn = db.connection()
+    dbapi = sa_conn.connection
+    driver = getattr(dbapi, "dbapi_connection", None) or getattr(
+        dbapi, "driver_connection", None
+    ) or dbapi
+    sql = (
+        "INSERT INTO biz_state_metric_row "
+        "(id,batch_id,batch_command_id,task_id,ne_id,metric_id,seq,data_json,collected_at) "
+        "VALUES %s"
+    )
+    with driver.cursor() as cur:
+        execute_values(
+            cur,
+            sql,
+            rows,
+            template="(%s,%s,%s,%s,%s,%s,%s,%s::jsonb,%s)",
+            page_size=len(rows),
+        )
+    db.flush()
+    return len(rows)
+
+
+def _persist_metric_rows_from_spool(
+    db,
+    *,
+    batch: BizStateBatch,
+    cmd_row: BizStateBatchCommand,
+    metric_id: str,
+    records_rel_path: str,
+) -> int:
+    """Stream JSONL → DB in chunks (never loads full table into memory)."""
+    from .spool import iter_record_chunks
+
+    mid = str(metric_id or "").strip()
+    if not mid or not records_rel_path:
+        return 0
+    use_copy = _db_is_postgres(db)
+    n = 0
+    for chunk in iter_record_chunks(records_rel_path, chunk_size=_METRIC_CHUNK):
+        if use_copy:
+            try:
+                added = _persist_metric_chunk_copy(
+                    db,
+                    batch=batch,
+                    cmd_row=cmd_row,
+                    metric_id=mid,
+                    chunk=chunk,
+                    seq_start=n,
+                )
+            except Exception:
+                _log.exception("biz_state COPY failed; falling back to bulk")
+                use_copy = False
+                added = _persist_metric_chunk_bulk(
+                    db,
+                    batch=batch,
+                    cmd_row=cmd_row,
+                    metric_id=mid,
+                    chunk=chunk,
+                    seq_start=n,
+                )
+        else:
+            added = _persist_metric_chunk_bulk(
+                db,
+                batch=batch,
+                cmd_row=cmd_row,
+                metric_id=mid,
+                chunk=chunk,
+                seq_start=n,
+            )
+        n += added
+    return n
+
+
+def _persist_lldp_rows_from_spool(
+    db,
+    *,
+    batch: BizStateBatch,
+    cmd_row: BizStateBatchCommand,
+    records_rel_path: str,
+) -> int:
+    from .spool import iter_record_chunks
+
+    if not records_rel_path:
+        return 0
+    n = 0
+    for chunk in iter_record_chunks(records_rel_path, chunk_size=_METRIC_CHUNK):
+        n += _persist_lldp_rows(db, batch=batch, cmd_row=cmd_row, records=chunk)
     return n
 
 
@@ -216,7 +409,6 @@ _GENERIC_METRICS = {
     "config_ospf",
     "config_isis",
 }
-_METRIC_CHUNK = 2000
 
 
 def _run_primary_parse_job(job: Any) -> tuple[bool, bool]:
@@ -239,10 +431,11 @@ def _run_primary_parse_job(job: Any) -> tuple[bool, bool]:
         m = re.search(r"(?i)total\s+number\s+of\s+routes\s*:\s*(\d+)", text)
         return int(m.group(1)) if m else 0
 
-    def _flush_item(item: SpooledCommand, *, records: list[dict[str, Any]] | None = None) -> None:
+    def _flush_item(item: SpooledCommand, *, records: Any = None) -> None:
         if records is not None and item.persist_kind:
-            item.records_rel_path = write_records(batch_id, item.id, records)
+            rel, n = write_records(batch_id, item.id, records)
-            item.row_count = len(records)
+            item.records_rel_path = rel
+            item.row_count = int(n or 0)
         try:
             write_meta(batch_id, item.id, item.to_meta())
         except Exception:
@@ -400,21 +593,25 @@ def _run_primary_parse_job(job: Any) -> tuple[bool, bool]:
                 "enrich=" + ",".join(getattr(j, "from_aux", "") for j in job.enrich_joins)
             )
         declared = int(primary.declared_total or 0)
-        nrec = len(records or [])
-        if declared > 0:
-            hints.append(f"declared={declared};parsed={nrec}")
-        if hints:
-            primary.message = ";".join(hints)[:1020]
         primary.parse_status = "ok"
         any_ok = True
         persist_recs = None
         if job.metric_id == "lldp_neighbor":
             primary.persist_kind = "lldp"
-            persist_recs = list(records or [])
+            persist_recs = records
         elif job.metric_id in _GENERIC_METRICS:
             primary.persist_kind = "metric"
-            persist_recs = list(records or [])
+            persist_recs = records
         _flush_item(primary, records=persist_recs)
+        nrec = int(primary.row_count or 0)
+        if declared > 0:
+            hints.append(f"declared={declared};parsed={nrec}")
+        if hints:
+            primary.message = ";".join(hints)[:1020]
+            try:
+                write_meta(batch_id, primary.id, primary.to_meta())
+            except Exception:
+                pass
         _ = fsm_tables  # kept for hints above
     except Exception as exc:
         any_fail = True
@@ -450,7 +647,7 @@ def _flush_spooled_commands(
     pending: list[Any],
 ) -> tuple[int, int]:
     """Insert SpooledCommand rows (+ metric/lldp) in one transaction. Returns (cmds, rows)."""
-    from .spool import SpooledCommand, raw_max_bytes, read_raw_text, read_records
+    from .spool import SpooledCommand, raw_max_bytes, read_raw_text, spool_file_size
 
     if not pending:
         return 0, 0
@@ -467,25 +664,38 @@ def _flush_spooled_commands(
             raw = ""
             truncated = False
             line_count = int(getattr(item, "raw_line_count", 0) or 0)
+            raw_size = 0
             if item.raw_rel_path:
                 from .spool import count_file_lines
 
+                raw_size = spool_file_size(item.raw_rel_path)
                 if line_count <= 0:
                     try:
                         line_count = count_file_lines(item.raw_rel_path)
                     except Exception:
                         line_count = 0
+                # Mega outputs stay on spool only — do not load full CLI into Postgres.
+                if max_raw > 0 and raw_size > max_raw:
+                    truncated = True
+                    raw = (
+                        f"[raw_on_spool={item.raw_rel_path}; size={raw_size}B; "
+                        f"cap={max_raw}B; omitted_from_db]\n"
+                    )
+                else:
                     raw = read_raw_text(item.raw_rel_path, max_bytes=max_raw)
                     if max_raw > 0 and "[truncated" in raw:
                         truncated = True
-            # Prefer full-file line count; fall back to stored text.
+            if line_count <= 0 and raw and not truncated:
-            if line_count <= 0 and raw:
                 from .spool import count_text_lines
 
                 line_count = count_text_lines(raw)
             msg = str(item.message or "").strip()
             if truncated:
-                note = f"raw_truncated@{max_raw}B"
+                note = (
+                    f"raw_on_spool@{raw_size}B"
+                    if raw_size > max_raw > 0
+                    else f"raw_truncated@{max_raw}B"
+                )
                 msg = f"{msg}; {note}" if msg else note
             cmd_row = BizStateBatchCommand(
                 id=item.id,
@@ -506,23 +716,23 @@ def _flush_spooled_commands(
             )
             db.add(cmd_row)
             if item.persist_kind == "metric" and item.records_rel_path:
-                records = read_records(item.records_rel_path)
                 mid = str(item.metric_id or "").strip()
-                if mid and records:
+                if mid:
-                    n = _persist_metric_rows(
+                    n = _persist_metric_rows_from_spool(
                         db,
                         batch=batch,
                         cmd_row=cmd_row,
                         metric_id=mid,
-                        records=records,
+                        records_rel_path=item.records_rel_path,
                     )
                     cmd_row.row_count = n
                     rows_n += n
             elif item.persist_kind == "lldp" and item.records_rel_path:
-                records = read_records(item.records_rel_path)
+                n = _persist_lldp_rows_from_spool(
-                if records:
+                    db,
-                    n = _persist_lldp_rows(
+                    batch=batch,
-                        db, batch=batch, cmd_row=cmd_row, records=records
+                    cmd_row=cmd_row,
+                    records_rel_path=item.records_rel_path,
                 )
                 cmd_row.row_count = n
                 rows_n += n
@@ -541,45 +751,6 @@ def _flush_spooled_commands(
         raise
 
 
-def _persist_metric_rows(
-    db,
-    *,
-    batch: BizStateBatch,
-    cmd_row: BizStateBatchCommand,
-    metric_id: str,
-    records: list[dict[str, Any]],
-) -> int:
-    """Bulk-insert generic metric rows (JSON payload per row)."""
-    mid = str(metric_id or "").strip()
-    if not mid or not records:
-        return 0
-    buf: list[dict[str, Any]] = []
-    n = 0
-    for i, rec in enumerate(records):
-        if not isinstance(rec, dict) or not rec:
-            continue
-        buf.append(
-            {
-                "id": uuid4().hex,
-                "batch_id": batch.id,
-                "batch_command_id": cmd_row.id,
-                "task_id": batch.task_id,
-                "ne_id": batch.ne_id,
-                "metric_id": mid,
-                "seq": i,
-                "data_json": dict(rec),
-                "collected_at": _utcnow(),
-            }
-        )
-        n += 1
-        if len(buf) >= _METRIC_CHUNK:
-            db.bulk_insert_mappings(BizStateMetricRow, buf)
-            buf.clear()
-    if buf:
-        db.bulk_insert_mappings(BizStateMetricRow, buf)
-    return n
-
-
 def _finish_task(task_id: str, *, error: str = "") -> None:
     db = SessionLocal()
     try:
@@ -825,11 +996,12 @@ def _run_collect_lane(
             pending = []
             persist.submit(batch_id, chunk)
 
-        def _queue(item: SpooledCommand, *, records: list[dict[str, Any]] | None = None) -> None:
+        def _queue(item: SpooledCommand, *, records: Any = None) -> None:
             nonlocal cmd_count
             if records is not None and item.persist_kind:
-                item.records_rel_path = write_records(batch_id, item.id, records)
+                rel, n = write_records(batch_id, item.id, records)
-                item.row_count = len(records)
+                item.records_rel_path = rel
+                item.row_count = int(n or 0)
             try:
                 write_meta(batch_id, item.id, item.to_meta())
             except Exception:

netx_api/biz_state/collect_session.py

@@ -341,7 +341,7 @@ def run_primary_with_bundle(
     textfsm_command: str = "",
     params: dict[str, str] | None = None,
     enrich_joins: list[EnrichJoin] | None = None,
-) -> tuple[list[dict[str, Any]], dict[str, list[dict[str, Any]]], list[str]]:
+) -> tuple[Any, dict[str, list[dict[str, Any]]], list[str]]:
     records, fsm_tables, keys = run_parser(
         parser_id,
         raw_text=bundle.raws.get("primary") or "",
@@ -356,5 +356,8 @@ def run_primary_with_bundle(
         fsm_tables_extra=bundle.fsm_extra,
     )
     if enrich_joins:
+        # Enrich needs random access — materialize only when joins are declared.
+        if not isinstance(records, list):
+            records = list(records or [])
         apply_enrich_joins(records, bundle.aux_records, enrich_joins)
     return records, fsm_tables, keys

netx_api/biz_state/parsers/__init__.py

@@ -34,6 +34,7 @@ Complex joins that cannot be expressed as equal-field copy still go in
 from __future__ import annotations
 
 import inspect
+from collections.abc import Iterable
 from typing import Any, Callable, Mapping, Sequence
 
 from ...lldp_shared import resolve_vendor_key
@@ -231,6 +232,11 @@ def run_parser(
             "params": params or {},
         }
     records = fn(**filtered)
-    if not isinstance(records, list):
+    # Allow list or streaming Iterable (generator); reject bare str/bytes.
+    if records is None:
+        records = []
+    elif isinstance(records, (str, bytes)):
+        records = []
+    elif not isinstance(records, list) and not isinstance(records, Iterable):
         records = []
     return records, fsm_tables, used_keys

netx_api/biz_state/parsers/zte/bgp_route.py

@@ -205,7 +205,6 @@ def _skip_noise_line(line: str) -> bool:
 
 
 def _emit_route(
-    out: list[dict[str, Any]],
     seen: set[str],
     *,
     local_as: str,
@@ -219,20 +218,19 @@ def _emit_route(
     rest: str,
     flags: str,
     path_continuation: bool = False,
-) -> None:
+) -> dict[str, Any] | None:
     if not _looks_like_prefix(net):
-        return
+        return None
     if nh and not _looks_like_ip_or_prefix(nh):
         if re.search(r"[A-Za-z]", nh):
-            return
+            return None
         nh = ""
     key = _route_dedupe_key(rd=rd, net=net, nh=nh or "")
     if key in seen:
-        return
+        return None
     seen.add(key)
     metric, loc, tag, path = _split_rest(rest, path_continuation=path_continuation)
-    out.append(
+    return {
-        {
         "local_as": local_as[:16],
         "afi": afi[:32],
         "vrf": vrf[:128],
@@ -250,7 +248,6 @@ def _emit_route(
         "state": "",
         "pfx_rcd": "",
     }
-    )
 
 
 def _hand_parse(
@@ -262,9 +259,8 @@ def _hand_parse(
     neighbor: str = "",
     direction: str = "",
     **_kw: Any,
-) -> list[dict[str, Any]]:
+):
-    """Parse neighbor in/out tables; join heavy IPv6 / From / metric wraps."""
+    """Yield neighbor in/out route rows (streaming; joins heavy IPv6 / From wraps)."""
-    out: list[dict[str, Any]] = []
     seen: set[str] = set()
     pending_net = ""
     pending_flags = ""
@@ -273,13 +269,12 @@ def _hand_parse(
     current_rd = ""
     current_vrf = vrf
 
-    def _flush_pending(*, rest: str = "", path_continuation: bool = False) -> None:
+    def _flush_pending(*, rest: str = "", path_continuation: bool = False) -> dict[str, Any] | None:
         nonlocal pending_net, pending_flags, pending_nh, pending_rest
         if not pending_net:
-            return
+            return None
         merged = " ".join(x for x in (pending_rest, rest) if x).strip()
-        _emit_route(
+        row = _emit_route(
-            out,
             seen,
             local_as=local_as,
             afi=afi,
@@ -297,20 +292,24 @@ def _hand_parse(
         pending_flags = ""
         pending_nh = ""
         pending_rest = ""
+        return row
 
-    def _start_pending(*, net: str, nh: str, flags: str, rest: str) -> None:
+    def _start_pending(*, net: str, nh: str, flags: str, rest: str):
         nonlocal pending_net, pending_flags, pending_nh, pending_rest
-        _flush_pending()
+        flushed = _flush_pending()
         pending_net = net
         pending_flags = flags
         pending_nh = nh
         pending_rest = rest
+        return flushed
 
     for raw in _normalize_cli_text(raw_text).splitlines():
         line = raw.rstrip()
         rd_m = _RD_RE.match(line.strip())
         if rd_m:
-            _flush_pending()
+            row = _flush_pending()
+            if row:
+                yield row
             current_rd = (rd_m.group("rd") or "").strip()
             vrf_from_rd = (rd_m.group("vrf") or "").strip()
             if vrf_from_rd:
@@ -331,15 +330,21 @@ def _hand_parse(
                     pending_nh = first
                     more = " ".join(parts[1:])
                     if more and _rest_looks_complete(more):
-                        _flush_pending(rest=more, path_continuation=True)
+                        row = _flush_pending(rest=more, path_continuation=True)
+                        if row:
+                            yield row
                     elif more:
                         pending_rest = " ".join(x for x in (pending_rest, more) if x)
                     continue
                 if not _looks_like_prefix(first):
-                    _flush_pending(rest=tok, path_continuation=True)
+                    row = _flush_pending(rest=tok, path_continuation=True)
+                    if row:
+                        yield row
                     continue
             else:
-                _flush_pending(rest=tok, path_continuation=True)
+                row = _flush_pending(rest=tok, path_continuation=True)
+                if row:
+                    yield row
                 continue
 
         # Peel optional status codes (* i / *i / > …)
@@ -355,9 +360,10 @@ def _hand_parse(
         if m and _looks_like_prefix(m.group("net")) and _looks_like_ip_or_prefix(m.group("nh")):
             rest = (m.group("rest") or "").strip()
             if _rest_looks_complete(rest):
-                _flush_pending()
+                row = _flush_pending()
-                _emit_route(
+                if row:
-                    out,
+                    yield row
+                row = _emit_route(
                     seen,
                     local_as=local_as,
                     afi=afi,
@@ -370,19 +376,27 @@ def _hand_parse(
                     rest=rest,
                     flags=flags,
                 )
+                if row:
+                    yield row
             else:
-                # Empty rest or From-only → wait for metric/path wrap
+                row = _start_pending(
-                _start_pending(net=m.group("net"), nh=m.group("nh"), flags=flags, rest=rest)
+                    net=m.group("net"), nh=m.group("nh"), flags=flags, rest=rest
+                )
+                if row:
+                    yield row
             continue
 
         # Network alone → wait for next-hop wrap
         m_net = _NET_ONLY_RE.match(body)
         if m_net and _looks_like_prefix(m_net.group("net")):
-            _start_pending(net=m_net.group("net"), nh="", flags=flags, rest="")
+            row = _start_pending(net=m_net.group("net"), nh="", flags=flags, rest="")
+            if row:
+                yield row
             continue
 
-    _flush_pending()
+    row = _flush_pending()
-    return out
+    if row:
+        yield row
 
 
 def normalize_bgp_route(
@@ -391,8 +405,8 @@ def normalize_bgp_route(
     command: str = "",
     params: dict[str, str] | None = None,
     **_kw: Any,
-) -> list[dict[str, Any]]:
+):
-    """Hand-only normalize (see module docstring)."""
+    """Hand-only normalize; returns [] or a streaming iterator of route dicts."""
     raw_text = _normalize_cli_text(raw_text)
     if _empty_if_total_zero(raw_text):
         return []

netx_api/biz_state/spool.py

@@ -6,6 +6,7 @@ import json
 import logging
 import re
 import shutil
+from collections.abc import Iterable, Iterator, Mapping
 from dataclasses import dataclass, field
 from pathlib import Path
 from typing import Any
@@ -78,15 +79,23 @@ def count_text_lines(text: str | None) -> int:
     return s.count("\n") + (0 if s.endswith("\n") else 1)
 
 
-def write_records(batch_id: str, cmd_id: str, records: list[dict[str, Any]]) -> str:
+def write_records(
-    """Write parsed records as JSONL; return relative path."""
+    batch_id: str,
+    cmd_id: str,
+    records: Iterable[Mapping[str, Any]] | None,
+) -> tuple[str, int]:
+    """Stream parsed records as JSONL; return (relative path, row count)."""
     _, _, rec_path = _cmd_paths(batch_id, cmd_id)
+    n = 0
     with rec_path.open("w", encoding="utf-8", errors="replace") as fh:
-        for rec in records or []:
+        for rec in records or ():
-            fh.write(json.dumps(rec, ensure_ascii=False, default=str))
+            if not isinstance(rec, Mapping):
+                continue
+            fh.write(json.dumps(dict(rec), ensure_ascii=False, default=str, separators=(",", ":")))
             fh.write("\n")
+            n += 1
     rel = rec_path.resolve().relative_to(spool_root())
-    return str(rel).replace("\\", "/")
+    return str(rel).replace("\\", "/"), n
 
 
 def write_meta(batch_id: str, cmd_id: str, meta: dict[str, Any]) -> str:
@@ -114,13 +123,26 @@ def read_raw_text(rel_path: str, *, max_bytes: int = 0) -> str:
     return data.decode("utf-8", errors="replace")
 
 
-def read_records(rel_path: str) -> list[dict[str, Any]]:
+def spool_file_size(rel_path: str) -> int:
+    """Byte size of a spool file; 0 if missing."""
     if not rel_path:
-        return []
+        return 0
     path = (spool_root() / str(rel_path)).resolve()
     if not str(path).startswith(str(spool_root())) or not path.is_file():
-        return []
+        return 0
-    out: list[dict[str, Any]] = []
+    try:
+        return int(path.stat().st_size)
+    except OSError:
+        return 0
+
+
+def iter_records(rel_path: str) -> Iterator[dict[str, Any]]:
+    """Yield one record dict at a time from JSONL (never loads full file)."""
+    if not rel_path:
+        return
+    path = (spool_root() / str(rel_path)).resolve()
+    if not str(path).startswith(str(spool_root())) or not path.is_file():
+        return
     with path.open("r", encoding="utf-8", errors="replace") as fh:
         for line in fh:
             line = line.strip()
@@ -131,8 +153,27 @@ def read_records(rel_path: str) -> list[dict[str, Any]]:
             except json.JSONDecodeError:
                 continue
             if isinstance(rec, dict):
-                out.append(rec)
+                yield rec
-    return out
+
+
+def iter_record_chunks(
+    rel_path: str, *, chunk_size: int = 2000
+) -> Iterator[list[dict[str, Any]]]:
+    """Yield lists of up to ``chunk_size`` records from JSONL."""
+    size = max(1, int(chunk_size or 2000))
+    buf: list[dict[str, Any]] = []
+    for rec in iter_records(rel_path):
+        buf.append(rec)
+        if len(buf) >= size:
+            yield buf
+            buf = []
+    if buf:
+        yield buf
+
+
+def read_records(rel_path: str) -> list[dict[str, Any]]:
+    """Load all records (tests / small payloads only — prefer iter_record_chunks)."""
+    return list(iter_records(rel_path))
 
 
 @dataclass

tests/test_biz_state_spool.py

@@ -17,6 +17,7 @@ from netx_api.biz_state import spool as spool_mod
 from netx_api.biz_state.spool import (
     SpooledCommand,
     clear_batch_spool,
+    iter_record_chunks,
     read_raw_text,
     read_records,
     write_raw_text,
@@ -43,11 +44,23 @@ class BizStateSpoolIoTests(unittest.TestCase):
         rel = write_raw_text(bid, cid, "show arp\nA B C")
         self.assertTrue(rel.endswith("cmd1.raw.txt"))
         self.assertEqual(read_raw_text(rel), "show arp\nA B C")
-        rrel = write_records(bid, cid, [{"ip": "1.1.1.1"}, {"ip": "2.2.2.2"}])
+        rrel, n = write_records(bid, cid, [{"ip": "1.1.1.1"}, {"ip": "2.2.2.2"}])
+        self.assertEqual(n, 2)
         recs = read_records(rrel)
         self.assertEqual(len(recs), 2)
         self.assertEqual(recs[0]["ip"], "1.1.1.1")
 
+    def test_write_records_streams_generator(self) -> None:
+        def _gen():
+            for i in range(5):
+                yield {"i": i}
+
+        rrel, n = write_records("batch-g", "cmd-g", _gen())
+        self.assertEqual(n, 5)
+        chunks = list(iter_record_chunks(rrel, chunk_size=2))
+        self.assertEqual([len(c) for c in chunks], [2, 2, 1])
+        self.assertEqual(chunks[0][0]["i"], 0)
+
     def test_raw_max_bytes_truncate(self) -> None:
         bid = "b2"
         cid = "c2"
@@ -116,11 +129,12 @@ class BizStateFlushSpoolTests(unittest.TestCase):
     def test_flush_inserts_command_and_metric_rows(self) -> None:
         cid = uuid4().hex
         raw_rel = write_raw_text("b-spool", cid, "ARP OUTPUT")
-        rec_rel = write_records(
+        rec_rel, rec_n = write_records(
             "b-spool",
             cid,
             [{"ip": "10.0.0.1", "mac": "aaaa"}, {"ip": "10.0.0.2", "mac": "bbbb"}],
         )
+        self.assertEqual(rec_n, 2)
         pending = [
             SpooledCommand(
                 id=cid,
@@ -134,6 +148,7 @@ class BizStateFlushSpoolTests(unittest.TestCase):
                 message="spooled",
                 raw_rel_path=raw_rel,
                 records_rel_path=rec_rel,
+                row_count=rec_n,
                 persist_kind="metric",
             )
         ]
@@ -158,6 +173,36 @@ class BizStateFlushSpoolTests(unittest.TestCase):
         self.assertEqual(batch.command_count, 1)
         self.assertEqual(batch.row_count, 2)
 
+    def test_flush_skips_mega_raw_into_db(self) -> None:
+        cid = uuid4().hex
+        big = "X" * (9 * 1024 * 1024)
+        raw_rel = write_raw_text("b-spool", cid, big)
+        rec_rel, rec_n = write_records("b-spool", cid, [{"k": 1}])
+        pending = [
+            SpooledCommand(
+                id=cid,
+                batch_id="b-spool",
+                metric_id="arp",
+                raw_command="show arp",
+                parse_status="ok",
+                message="ok",
+                raw_rel_path=raw_rel,
+                records_rel_path=rec_rel,
+                row_count=rec_n,
+                persist_kind="metric",
+            )
+        ]
+        with patch.object(spool_mod.settings, "biz_state_raw_max_bytes", 8 * 1024 * 1024):
+            cmds, rows = runner._flush_spooled_commands("b-spool", pending)
+        self.assertEqual(cmds, 1)
+        self.assertEqual(rows, 1)
+        self.db.expire_all()
+        cmd = self.db.get(BizStateBatchCommand, cid)
+        assert cmd is not None
+        self.assertIn("raw_on_spool", cmd.raw_text)
+        self.assertNotIn("XXXX", cmd.raw_text)
+        self.assertLess(len(cmd.raw_text or ""), 500)
+
     def test_flush_batches_multiple_without_per_cmd_sessions(self) -> None:
         pending: list[SpooledCommand] = []
         for i in range(5):

tests/test_zte_extended_parsers.py

@@ -12,7 +12,7 @@ from netx_api.biz_state.enrich import apply_enrich_joins
 from netx_api.biz_state.parsers.common.vrf_list import normalize_vrf_list
 from netx_api.biz_state.parsers.zte import (
     normalize_bgp_peer,
-    normalize_bgp_route,
+    normalize_bgp_route as _normalize_bgp_route_stream,
     normalize_ip_route,
     normalize_ipv6_route,
     normalize_l2vpn_mac,
@@ -26,6 +26,12 @@ from netx_api.biz_state.profiles import AuxCommand, get_profile, metric_field_ma
 from netx_api.ntc_parse import apply_rule
 
 
+def normalize_bgp_route(**kwargs):
+    """Materialize streaming parser for assertions / enrich joins."""
+    out = _normalize_bgp_route_stream(**kwargs)
+    return out if isinstance(out, list) else list(out)
+
+
 _VRF_SAMPLE = """
 Name                             Default RD            Protocols VRF ID
 CUST_A                           100:1                 ipv4      1