Implement rcon-first historical ingestion path

2026-03-25 21:38:02 +01:00
parent 6ed416f79a
commit cdcd4523b4
6 changed files with 547 additions and 185 deletions
--- a/ai/tasks/done/TASK-100-rcon-historical-writer-path-implementation.md
+++ b/ai/tasks/done/TASK-100-rcon-historical-writer-path-implementation.md
@@ -0,0 +1,53 @@
+# TASK-100-rcon-historical-writer-path-implementation
+
+## Goal
+Implementar un writer path histórico real por RCON para que la ingesta histórica intente RCON primero y use scoreboard/public-scoreboard solo como fallback.
+
+## Context
+La repo ya tiene:
+- live RCON-first
+- captura prospectiva RCON
+- read model histórico RCON parcial
+- ingesta histórica clásica por scoreboard
+Pero todavía no existe un writer path histórico real por RCON integrado en `historical_ingestion`.
+
+## Steps
+1. Auditar:
+   - `backend/app/data_sources.py`
+   - `backend/app/historical_ingestion.py`
+   - `backend/app/rcon_historical_worker.py`
+   - `backend/app/rcon_historical_storage.py`
+   - `backend/app/rcon_historical_read_model.py`
+   - cualquier capa necesaria de modelos/storage
+2. Definir qué significa “writer path histórico RCON” con la telemetría real actual:
+   - qué puede alimentar
+   - qué estructura persistida necesita
+   - cómo se integra con el ingestion flow existente
+3. Implementar una vía writer-oriented RCON que permita a `historical_ingestion` intentar primero RCON.
+4. Si RCON falla o no cubre la operación concreta, hacer fallback controlado a scoreboard/public-scoreboard.
+5. Mantener trazabilidad explícita:
+   - primary_source
+   - selected_source
+   - fallback_used
+   - fallback_reason
+   - source_attempts
+6. No romper la compatibilidad con snapshots ni con rebuilds posteriores.
+7. Actualizar README/runbook explicando el nuevo writer path real.
+
+## Constraints
+- No fingir cobertura histórica RCON que no exista.
+- No eliminar scoreboard como fallback.
+- No romper live RCON-first.
+- No romper el locking compartido.
+
+## Validation
+- `historical_ingestion` intenta RCON primero.
+- Cuando RCON falla o no soporta la operación, hace fallback explícito a scoreboard.
+- La salida del comando deja claro qué fuente se usó realmente.
+- El repositorio queda consistente.
+
+## Expected Files
+- `backend/app/data_sources.py`
+- `backend/app/historical_ingestion.py`
+- archivos backend necesarios para writer path RCON
+- `backend/README.md`
--- a/ai/tasks/done/TASK-101-remove-false-rcon-first-claims-and-fix-operational-visibility.md
+++ b/ai/tasks/done/TASK-101-remove-false-rcon-first-claims-and-fix-operational-visibility.md
@@ -0,0 +1,38 @@
+# TASK-101-remove-false-rcon-first-claims-and-fix-operational-visibility
+
+## Goal
+Alinear documentación, outputs operativos y visibilidad de progreso con el comportamiento real del sistema, evitando afirmaciones engañosas sobre histórico RCON-first y mejorando la operativa del refresh manual.
+
+## Context
+Actualmente el operador puede pensar que la ingesta histórica ya va por RCON cuando en realidad el writer path sigue cayendo a scoreboard.
+Además, el comando de refresh es demasiado opaco: tarda mucho y no ofrece progreso útil.
+
+## Steps
+1. Auditar:
+   - `backend/README.md`
+   - `backend/app/historical_ingestion.py`
+   - outputs/logs relevantes
+2. Corregir cualquier copy/documentación que sugiera que la ingesta histórica completa ya está en RCON si no es cierto.
+3. Añadir progreso operativo útil al refresh manual:
+   - servidor actual
+   - página actual
+   - número de match ids a detallar
+   - fuente realmente seleccionada
+4. Hacer que, cuando haya fallback a scoreboard, eso quede visible para el operador en tiempo real o en el payload final.
+5. Mantener la salida usable, sin inundar de logs innecesarios.
+6. Actualizar runbook con recomendaciones reales para pasadas manuales y límites razonables.
+
+## Constraints
+- No convertir el comando en un spam de logs.
+- No ocultar fallbacks reales.
+- No mezclar esta task con grandes cambios de UI frontend.
+
+## Validation
+- El operador puede ver progreso útil durante un refresh.
+- La fuente usada de verdad queda visible.
+- La documentación ya no induce a error.
+- El repositorio queda consistente.
+
+## Expected Files
+- `backend/app/historical_ingestion.py`
+- `backend/README.md`
--- a/backend/README.md
+++ b/backend/README.md
@@ -272,7 +272,7 @@ Valores soportados en esta fase:
  - `rcon` como camino primario recomendado
  - `a2s` como fallback legacy o override explicito
 - historico:
-  - `rcon` como camino primario recomendado para captura y lectura minima
+  - `rcon` como camino primario recomendado para captura y writer path primario
  - `public-scoreboard` como fallback legacy o override explicito

 Defaults actuales:
@@ -309,18 +309,14 @@ Politica funcional actual:
  - `public-scoreboard` solo si RCON no soporta aun esa operacion concreta o
    falla la captura primaria

-Limitacion actual de `rcon`:
-
- el backend puede usar `rcon` para `/api/servers`
- la ingesta historica competitiva por `historical_ingestion.py` sigue cayendo a
-  `public-scoreboard`, porque la repo todavia no incluye una canalizacion
-  retroactiva RCON capaz de reconstruir partidas cerradas con paridad completa
-
 Estado real de "historico por RCON" en esta repo:

 - no existe backfill retroactivo por RCON con el cliente actual
 - la viabilidad documentada hoy es solo para captura prospectiva separada
- `public-scoreboard` sigue siendo la fuente historica principal
+- `historical_ingestion.py` intenta primero el writer path prospectivo RCON
+- la persistencia competitiva `historical_*` sigue necesitando fallback a
+  `public-scoreboard` mientras RCON no exponga pagina historica/detalle de
+  match cerrada con paridad suficiente
 - el diseno tecnico de esa linea prospectiva queda en
  `docs/rcon-historical-ingestion-design.md`

@@ -473,7 +469,8 @@ respuesta deja trazabilidad con `primary_source`, `selected_source`,
  consultas live de produccion.
 - `config.py` centraliza host, puerto y allowlist minima de origenes locales.
 - `data_sources.py` define los contratos y la seleccion por entorno para live e historico.
- `historical_ingestion.py` consulta la capa JSON publica de CRCON para bootstrap y refresh incremental.
+- `historical_ingestion.py` intenta primero el writer path RCON y, si hace falta
+  poblar `historical_*`, cae de forma explicita a la capa JSON publica de CRCON.
 - `historical_models.py` fija las entidades historicas minimas del dominio.
 - `historical_snapshots.py` fija los tipos y selectores validos de snapshots historicos precalculados.
 - `historical_snapshot_storage.py` persiste snapshots historicos precalculados listos para lectura rapida.
@@ -1039,6 +1036,25 @@ Si la recomposicion se lanza para un servidor fisico concreto, el backend
 rehace tambien el agregado logico `all-servers` para mantener `Todos`
 alineado con `#01` y `#02` aunque `#03` siga sin bootstrap.

+En esta fase, el comando muestra progreso operativo util sin saturar stdout:
+
+- intento primario RCON
+- fuente finalmente seleccionada
+- servidor actual
+- pagina actual
+- `match_ids_to_detail` de cada pagina
+
+Si RCON no puede cubrir la operacion competitiva real, el fallback a
+`public-scoreboard` queda visible tanto durante la ejecucion como en el JSON
+final mediante:
+
+- `primary_source`
+- `selected_source`
+- `fallback_used`
+- `fallback_reason`
+- `source_attempts`
+- `primary_writer_result`
+
 El comando devuelve ademas un resumen de cobertura persistida por servidor. Esto
 ayuda a validar rapidamente cuantos matches reales quedaron importados, el rango
 temporal cubierto y si la carga ya supera la ultima semana movil que usa la UI.
@@ -1080,6 +1096,16 @@ registrados. La segunda sirve para validar un solo servidor con alcance
 acotado. La tercera recompone snapshots despues de una pasada manual cuando se
 quiere confirmar que la capa precalculada vuelve a quedar alineada.

+Interpretacion operativa recomendada:
+
+- si aparece `historical-ingestion-rcon-primary-succeeded`, RCON se intento de
+  verdad primero y la captura prospectiva quedo registrada
+- si despues `selected_source` termina en `public-scoreboard`, eso significa
+  que la reconstruccion del archivo competitivo `historical_*` siguio
+  necesitando fallback clasico
+- si RCON falla por red, auth o timeout, el motivo queda visible en
+  `fallback_reason` y en `source_attempts`
+
 Los reintentos de cada request JSON pueden ajustarse sin tocar codigo con:

 - `HLL_HISTORICAL_CRCON_REQUEST_RETRIES`
@@ -1392,9 +1418,11 @@ El backend queda orientado a `RCON-first` tambien para historico:
  - `rcon` primero
  - `a2s` solo como fallback
 - historico:
-  - `rcon` primero para la capa de lectura/cobertura soportada hoy
+  - `rcon` primero tanto para lectura minima como para el writer path primario
+    de `historical_ingestion`
  - `public-scoreboard` solo como fallback cuando RCON no cubre una operacion
-    competitiva concreta o no tiene cobertura suficiente
+    competitiva concreta, no tiene cobertura suficiente o falla la captura
+    primaria

 Metadata observable en payloads historicos:

@@ -1407,6 +1435,8 @@ Metadata observable en payloads historicos:
 Estado real a fecha de esta fase:

 - el read model historico RCON soporta cobertura y actividad reciente
+- `historical_ingestion` intenta primero una captura writer-oriented por RCON y
+  deja esa tentativa visible en su salida
 - rankings competitivos, MVP y Elo/MMR siguen necesitando fallback a
  `public-scoreboard` porque el read model RCON actual no expone aun detalle
  historico competitivo suficiente
--- a/backend/app/data_sources.py
+++ b/backend/app/data_sources.py
@@ -327,45 +327,46 @@ def build_historical_runtime_source_policy(


 def resolve_historical_ingestion_data_source() -> tuple[HistoricalDataSource, dict[str, object]]:
-    """Resolve the writer-oriented historical provider with safe fallback semantics."""
+    """Resolve the fallback provider used when classic scoreboard import is required."""
    configured_kind = get_historical_data_source_kind()
-    if configured_kind == SOURCE_KIND_PUBLIC_SCOREBOARD:
-        return (
-            PublicScoreboardHistoricalDataSource(),
-            build_source_policy(
-                primary_source=SOURCE_KIND_PUBLIC_SCOREBOARD,
-                selected_source=SOURCE_KIND_PUBLIC_SCOREBOARD,
-                source_attempts=[
-                    build_source_attempt(
-                        source=SOURCE_KIND_PUBLIC_SCOREBOARD,
-                        role="primary",
-                        status="success",
-                    )
-                ],
-            ),
+    if configured_kind in {SOURCE_KIND_PUBLIC_SCOREBOARD, SOURCE_KIND_RCON}:
+        primary_source = (
+            SOURCE_KIND_PUBLIC_SCOREBOARD
+            if configured_kind == SOURCE_KIND_PUBLIC_SCOREBOARD
+            else SOURCE_KIND_RCON
+        )
+        fallback_used = configured_kind == SOURCE_KIND_RCON
+        fallback_reason = (
+            "classic-historical-import-requires-public-scoreboard-fallback"
+            if fallback_used
+            else None
+        )
+        attempts = []
+        if configured_kind == SOURCE_KIND_RCON:
+            attempts.append(
+                build_source_attempt(
+                    source=SOURCE_KIND_RCON,
+                    role="primary",
+                    status="deferred",
+                    reason="rcon-primary-writer-attempt-is-handled-by-historical-ingestion",
+                )
+            )
+        attempts.append(
+            build_source_attempt(
+                source=SOURCE_KIND_PUBLIC_SCOREBOARD,
+                role="fallback" if fallback_used else "primary",
+                status="ready",
+                reason="classic-historical-import-provider-ready",
+            )
        )
-
-    if configured_kind == SOURCE_KIND_RCON:
        return (
            PublicScoreboardHistoricalDataSource(),
            build_source_policy(
-                primary_source=SOURCE_KIND_RCON,
+                primary_source=primary_source,
                selected_source=SOURCE_KIND_PUBLIC_SCOREBOARD,
-                fallback_used=True,
-                fallback_reason="rcon-historical-ingestion-not-supported-yet",
-                source_attempts=[
-                    build_source_attempt(
-                        source=SOURCE_KIND_RCON,
-                        role="primary",
-                        status="unsupported",
-                        reason="rcon-historical-ingestion-not-supported-yet",
-                    ),
-                    build_source_attempt(
-                        source=SOURCE_KIND_PUBLIC_SCOREBOARD,
-                        role="fallback",
-                        status="success",
-                    ),
-                ],
+                fallback_used=fallback_used,
+                fallback_reason=fallback_reason,
+                source_attempts=attempts,
            ),
        )

--- a/backend/app/historical_ingestion.py
+++ b/backend/app/historical_ingestion.py
@@ -5,14 +5,21 @@ from __future__ import annotations
 import argparse
 import json
 from dataclasses import dataclass
-from typing import Iterable
+from typing import Callable, Iterable

 from .config import (
    get_historical_crcon_detail_workers,
    get_historical_crcon_page_size,
+    get_historical_data_source_kind,
    get_historical_refresh_overlap_hours,
 )
-from .data_sources import HistoricalDataSource, resolve_historical_ingestion_data_source
+from .data_sources import (
+    SOURCE_KIND_PUBLIC_SCOREBOARD,
+    SOURCE_KIND_RCON,
+    HistoricalDataSource,
+    build_historical_runtime_source_policy,
+    resolve_historical_ingestion_data_source,
+)
 from .elo_mmr_engine import rebuild_elo_mmr_models
 from .historical_snapshots import generate_and_persist_historical_snapshots
 from .historical_storage import (
@@ -28,9 +35,13 @@ from .historical_storage import (
    start_ingestion_run,
    upsert_historical_match,
 )
+from .rcon_historical_worker import run_rcon_historical_capture_unlocked
 from .writer_lock import backend_writer_lock, build_writer_lock_holder


+ProgressCallback = Callable[[dict[str, object]], None]
+
+
@dataclass(slots=True)
 class IngestionStats:
    """Mutable counters for one ingestion execution."""
@@ -57,6 +68,7 @@ def run_bootstrap(
    start_page: int | None = None,
    detail_workers: int | None = None,
    rebuild_snapshots: bool = True,
+    progress_callback: ProgressCallback | None = None,
 ) -> dict[str, object]:
    """Run a first full historical import against one or all configured servers."""
    with backend_writer_lock(
@@ -74,6 +86,7 @@ def run_bootstrap(
            overlap_hours=None,
            incremental=False,
            rebuild_snapshots=rebuild_snapshots,
+            progress_callback=progress_callback,
        )


@@ -86,6 +99,7 @@ def run_incremental_refresh(
    detail_workers: int | None = None,
    overlap_hours: int | None = None,
    rebuild_snapshots: bool = True,
+    progress_callback: ProgressCallback | None = None,
 ) -> dict[str, object]:
    """Refresh recent historical pages without replaying the whole archive."""
    with backend_writer_lock(
@@ -103,6 +117,7 @@ def run_incremental_refresh(
            overlap_hours=overlap_hours,
            incremental=True,
            rebuild_snapshots=rebuild_snapshots,
+            progress_callback=progress_callback,
        )


@@ -117,10 +132,11 @@ def _run_ingestion(
    overlap_hours: int | None,
    incremental: bool,
    rebuild_snapshots: bool,
+    progress_callback: ProgressCallback | None,
 ) -> dict[str, object]:
    initialize_historical_storage()
    stats = IngestionStats()
-    data_source, source_policy = resolve_historical_ingestion_data_source()
+    fallback_data_source, fallback_source_policy = resolve_historical_ingestion_data_source()
    selected_servers = _select_servers(server_slug)
    processed_servers: list[dict[str, object]] = []
    active_runs: dict[str, int] = {}
@@ -132,60 +148,86 @@ def _run_ingestion(
    if resolved_overlap_hours < 0:
        raise ValueError("--overlap-hours must be zero or positive.")

+    primary_writer_result = _attempt_primary_rcon_writer(
+        mode=mode,
+        server_slug=server_slug,
+        selected_servers=selected_servers,
+        progress_callback=progress_callback,
+    )
+    source_policy = _resolve_ingestion_source_policy(
+        fallback_source_policy=fallback_source_policy,
+        primary_writer_result=primary_writer_result,
+    )
+    use_classic_fallback = _should_use_classic_fallback(primary_writer_result)
+    _emit_progress(
+        progress_callback,
+        {
+            "event": "historical-ingestion-source-selected",
+            "mode": mode,
+            "primary_source": source_policy.get("primary_source"),
+            "selected_source": source_policy.get("selected_source"),
+            "fallback_used": bool(source_policy.get("fallback_used")),
+            "fallback_reason": source_policy.get("fallback_reason"),
+        },
+    )
+
    try:
-        for server in selected_servers:
-            run_id = start_ingestion_run(mode=mode, target_server_slug=str(server["slug"]))
-            active_runs[str(server["slug"])] = run_id
-            mark_backfill_progress_started(
-                server_slug=str(server["slug"]),
-                mode=mode,
-                run_id=run_id,
-            )
-            cutoff = (
-                get_refresh_cutoff_for_server(
-                    str(server["slug"]),
-                    overlap_hours=resolved_overlap_hours,
+        if use_classic_fallback:
+            for server in selected_servers:
+                run_id = start_ingestion_run(mode=mode, target_server_slug=str(server["slug"]))
+                active_runs[str(server["slug"])] = run_id
+                mark_backfill_progress_started(
+                    server_slug=str(server["slug"]),
+                    mode=mode,
+                    run_id=run_id,
                )
-                if incremental
-                else None
-            )
-            resolved_start_page = _resolve_start_page(
-                start_page=start_page,
-                server_slug=str(server["slug"]),
-                mode=mode,
-            )
-            server_stats = _ingest_server(
-                server=server,
-                mode=mode,
-                run_id=run_id,
-                stats=stats,
-                data_source=data_source,
-                max_pages=max_pages,
-                page_size=page_size,
-                start_page=resolved_start_page,
-                detail_workers=detail_workers,
-                cutoff=cutoff,
-            )
-            processed_servers.append(server_stats)
-            finalize_ingestion_run(
-                run_id,
-                status="success",
-                pages_processed=server_stats["pages_processed"],
-                matches_seen=server_stats["matches_seen"],
-                matches_inserted=server_stats["matches_inserted"],
-                matches_updated=server_stats["matches_updated"],
-                player_rows_inserted=server_stats["player_rows_inserted"],
-                player_rows_updated=server_stats["player_rows_updated"],
-                notes=f"public_name={server_stats['public_name']}",
-            )
-            finalize_backfill_progress(
-                server_slug=str(server["slug"]),
-                mode=mode,
-                run_id=run_id,
-                status="success",
-                archive_exhausted=bool(server_stats["archive_exhausted"]),
-            )
-            active_runs.pop(str(server["slug"]), None)
+                cutoff = (
+                    get_refresh_cutoff_for_server(
+                        str(server["slug"]),
+                        overlap_hours=resolved_overlap_hours,
+                    )
+                    if incremental
+                    else None
+                )
+                resolved_start_page = _resolve_start_page(
+                    start_page=start_page,
+                    server_slug=str(server["slug"]),
+                    mode=mode,
+                )
+                server_stats = _ingest_server(
+                    server=server,
+                    mode=mode,
+                    run_id=run_id,
+                    stats=stats,
+                    data_source=fallback_data_source,
+                    max_pages=max_pages,
+                    page_size=page_size,
+                    start_page=resolved_start_page,
+                    detail_workers=detail_workers,
+                    cutoff=cutoff,
+                    progress_callback=progress_callback,
+                    source_policy=source_policy,
+                )
+                processed_servers.append(server_stats)
+                finalize_ingestion_run(
+                    run_id,
+                    status="success",
+                    pages_processed=server_stats["pages_processed"],
+                    matches_seen=server_stats["matches_seen"],
+                    matches_inserted=server_stats["matches_inserted"],
+                    matches_updated=server_stats["matches_updated"],
+                    player_rows_inserted=server_stats["player_rows_inserted"],
+                    player_rows_updated=server_stats["player_rows_updated"],
+                    notes=f"public_name={server_stats['public_name']}",
+                )
+                finalize_backfill_progress(
+                    server_slug=str(server["slug"]),
+                    mode=mode,
+                    run_id=run_id,
+                    status="success",
+                    archive_exhausted=bool(server_stats["archive_exhausted"]),
+                )
+                active_runs.pop(str(server["slug"]), None)
        if rebuild_snapshots:
            snapshot_result = generate_and_persist_historical_snapshots(server_key=server_slug)
            elo_mmr_result = rebuild_elo_mmr_models()
@@ -224,8 +266,9 @@ def _run_ingestion(
    return {
        "status": "ok",
        "mode": mode,
-        "source_provider": data_source.source_kind,
+        "source_provider": source_policy.get("selected_source"),
        "source_policy": source_policy,
+        "primary_writer_result": primary_writer_result,
        "page_size": page_size or get_historical_crcon_page_size(),
        "start_page": start_page,
        "detail_workers": detail_workers or get_historical_crcon_detail_workers(),
@@ -257,6 +300,8 @@ def _ingest_server(
    start_page: int,
    detail_workers: int | None,
    cutoff: str | None,
+    progress_callback: ProgressCallback | None,
+    source_policy: dict[str, object],
 ) -> dict[str, object]:
    resolved_page_size = page_size or get_historical_crcon_page_size()
    resolved_detail_workers = detail_workers or get_historical_crcon_detail_workers()
@@ -267,6 +312,18 @@ def _ingest_server(
    discovered_total_matches: int | None = None
    last_page_processed: int | None = None
    archive_exhausted = False
+    _emit_progress(
+        progress_callback,
+        {
+            "event": "historical-ingestion-server-started",
+            "mode": mode,
+            "server_slug": server["slug"],
+            "selected_source": source_policy.get("selected_source"),
+            "fallback_used": bool(source_policy.get("fallback_used")),
+            "start_page": start_page,
+            "cutoff": cutoff,
+        },
+    )

    for page_number in range(start_page, start_page + page_limit):
        payload = data_source.fetch_match_page(
@@ -300,6 +357,20 @@ def _ingest_server(
            if match_id:
                match_ids_to_fetch.append(match_id)

+        _emit_progress(
+            progress_callback,
+            {
+                "event": "historical-ingestion-page-loaded",
+                "mode": mode,
+                "server_slug": server["slug"],
+                "page": page_number,
+                "selected_source": source_policy.get("selected_source"),
+                "match_ids_to_detail": len(match_ids_to_fetch),
+                "page_matches": len(page_matches),
+                "cutoff_reached": stop_after_page,
+            },
+        )
+
        for detail_payload in data_source.fetch_match_details(
            base_url=str(server["scoreboard_base_url"]),
            match_ids=match_ids_to_fetch,
@@ -356,6 +427,162 @@ def _resolve_start_page(
    return get_backfill_resume_page(server_slug, mode=mode)


+def _attempt_primary_rcon_writer(
+    *,
+    mode: str,
+    server_slug: str | None,
+    selected_servers: list[dict[str, object]],
+    progress_callback: ProgressCallback | None,
+) -> dict[str, object]:
+    configured_kind = get_historical_data_source_kind()
+    if configured_kind != SOURCE_KIND_RCON:
+        result = {
+            "attempted": False,
+            "status": "skipped",
+            "primary_source": SOURCE_KIND_PUBLIC_SCOREBOARD,
+            "selected_source": SOURCE_KIND_PUBLIC_SCOREBOARD,
+            "fallback_used": False,
+            "fallback_reason": None,
+            "source_attempts": [],
+        }
+        _emit_progress(
+            progress_callback,
+            {
+                "event": "historical-ingestion-rcon-primary-skipped",
+                "mode": mode,
+                "reason": "historical-data-source-configured-for-public-scoreboard",
+            },
+        )
+        return result
+
+    target_scope = server_slug or "all-configured-rcon-targets"
+    _emit_progress(
+        progress_callback,
+        {
+            "event": "historical-ingestion-rcon-primary-started",
+            "mode": mode,
+            "target_scope": target_scope,
+            "servers": [str(server["slug"]) for server in selected_servers],
+        },
+    )
+    try:
+        capture_result = run_rcon_historical_capture_unlocked(target_key=server_slug)
+    except Exception as exc:  # noqa: BLE001 - fallback remains explicit and controlled
+        result = {
+            "attempted": True,
+            "status": "error",
+            "primary_source": SOURCE_KIND_RCON,
+            "selected_source": SOURCE_KIND_PUBLIC_SCOREBOARD,
+            "fallback_used": True,
+            "fallback_reason": "rcon-historical-writer-request-failed",
+            "message": str(exc),
+        }
+        _emit_progress(
+            progress_callback,
+            {
+                "event": "historical-ingestion-rcon-primary-failed",
+                "mode": mode,
+                "target_scope": target_scope,
+                "message": str(exc),
+            },
+        )
+        return result
+
+    capture_run_status = str(capture_result.get("run_status") or capture_result.get("status") or "unknown")
+    targets = list(capture_result.get("targets") or [])
+    errors = list(capture_result.get("errors") or [])
+    if targets:
+        result = {
+            "attempted": True,
+            "status": "partial",
+            "primary_source": SOURCE_KIND_RCON,
+            "selected_source": SOURCE_KIND_PUBLIC_SCOREBOARD,
+            "fallback_used": True,
+            "fallback_reason": "rcon-primary-writer-succeeded-but-classic-match-archive-still-needs-fallback",
+            "capture_result": capture_result,
+        }
+        _emit_progress(
+            progress_callback,
+            {
+                "event": "historical-ingestion-rcon-primary-succeeded",
+                "mode": mode,
+                "target_scope": target_scope,
+                "captured_targets": len(targets),
+                "run_status": capture_run_status,
+                "next_step": "classic-public-scoreboard-fallback-required",
+            },
+        )
+        return result
+
+    result = {
+        "attempted": True,
+        "status": "empty",
+        "primary_source": SOURCE_KIND_RCON,
+        "selected_source": SOURCE_KIND_PUBLIC_SCOREBOARD,
+        "fallback_used": True,
+        "fallback_reason": "rcon-historical-writer-returned-no-usable-samples",
+        "capture_result": capture_result,
+        "message": json.dumps(errors, separators=(",", ":")) if errors else None,
+    }
+    _emit_progress(
+        progress_callback,
+        {
+            "event": "historical-ingestion-rcon-primary-empty",
+            "mode": mode,
+            "target_scope": target_scope,
+            "run_status": capture_run_status,
+            "errors": len(errors),
+        },
+    )
+    return result
+
+
+def _should_use_classic_fallback(primary_writer_result: dict[str, object]) -> bool:
+    selected_source = str(primary_writer_result.get("selected_source") or "")
+    return selected_source == SOURCE_KIND_PUBLIC_SCOREBOARD
+
+
+def _resolve_ingestion_source_policy(
+    *,
+    fallback_source_policy: dict[str, object],
+    primary_writer_result: dict[str, object],
+) -> dict[str, object]:
+    configured_kind = get_historical_data_source_kind()
+    if configured_kind != SOURCE_KIND_RCON:
+        return fallback_source_policy
+
+    status = str(primary_writer_result.get("status") or "error")
+    selected_source = str(
+        primary_writer_result.get("selected_source") or SOURCE_KIND_PUBLIC_SCOREBOARD
+    )
+    fallback_reason = primary_writer_result.get("fallback_reason")
+    message = primary_writer_result.get("message")
+    if (
+        fallback_reason
+        == "rcon-primary-writer-succeeded-but-classic-match-archive-still-needs-fallback"
+    ):
+        message = (
+            "RCON prospective capture succeeded first, but the classic historical_* "
+            "archive still requires public-scoreboard for match-page import."
+        )
+    return build_historical_runtime_source_policy(
+        operation="historical-ingestion",
+        rcon_status=status,
+        fallback_reason=str(fallback_reason) if fallback_reason else None,
+        selected_source=selected_source,
+        rcon_message=message if isinstance(message, str) else None,
+    )
+
+
+def _emit_progress(
+    callback: ProgressCallback | None,
+    payload: dict[str, object],
+) -> None:
+    if callback is None:
+        return
+    callback(payload)
+
+
 def _select_servers(server_slug: str | None) -> list[dict[str, object]]:
    servers = list_historical_servers()
    if server_slug is None:
@@ -456,6 +683,9 @@ def main(argv: Iterable[str] | None = None) -> int:
    parser = build_arg_parser()
    args = parser.parse_args(list(argv) if argv is not None else None)

+    def _print_progress(payload: dict[str, object]) -> None:
+        print(json.dumps(payload, ensure_ascii=True))
+
    if args.mode == "bootstrap":
        result = run_bootstrap(
            server_slug=args.server_slug,
@@ -463,6 +693,7 @@ def main(argv: Iterable[str] | None = None) -> int:
            page_size=args.page_size,
            start_page=args.start_page,
            detail_workers=args.detail_workers,
+            progress_callback=_print_progress,
        )
    else:
        result = run_incremental_refresh(
@@ -472,6 +703,7 @@ def main(argv: Iterable[str] | None = None) -> int:
            start_page=args.start_page,
            detail_workers=args.detail_workers,
            overlap_hours=args.overlap_hours,
+            progress_callback=_print_progress,
        )

    print(json.dumps(result, indent=2))
--- a/backend/app/rcon_historical_worker.py
+++ b/backend/app/rcon_historical_worker.py
@@ -44,93 +44,101 @@ def run_rcon_historical_capture(
            f"app.rcon_historical_worker capture:{target_key or 'all-targets'}"
        )
    ):
-        initialize_rcon_historical_storage()
-        selected_targets = _select_targets(target_key)
-        captured_at = utc_now().isoformat().replace("+00:00", "Z")
-        target_scope = target_key or "all-configured-rcon-targets"
-        run_id = start_rcon_historical_capture_run(mode="capture", target_scope=target_scope)
-        stats = RconHistoricalCaptureStats()
-        items: list[dict[str, object]] = []
-        errors: list[dict[str, object]] = []
+        return run_rcon_historical_capture_unlocked(target_key=target_key)

-        try:
-            for target in selected_targets:
-                target_metadata = _serialize_target(target)
-                stats.targets_seen += 1
-                try:
-                    sample = query_live_server_sample(target)
-                    delta = persist_rcon_historical_sample(
-                        run_id=run_id,
-                        captured_at=captured_at,
-                        target=target_metadata,
-                        normalized_payload=sample["normalized"],
-                        raw_payload=sample["raw_session"],
-                    )
-                    stats.samples_inserted += int(delta["samples_inserted"])
-                    stats.duplicate_samples += int(delta["duplicate_samples"])
-                    items.append(
-                        {
-                            "target_key": target_metadata["target_key"],
-                            "external_server_id": target.external_server_id,
-                            "captured_at": captured_at,
-                            "sample_inserted": bool(delta["samples_inserted"]),
-                            "normalized": sample["normalized"],
-                        }
-                    )
-                except Exception as exc:  # noqa: BLE001 - controlled worker failures
-                    stats.failed_targets += 1
-                    mark_rcon_historical_capture_failure(
-                        run_id=run_id,
-                        target=target_metadata,
-                        error_message=str(exc),
-                    )
-                    errors.append(
-                        {
-                            "target_key": target_metadata["target_key"],
-                            "name": target.name,
-                            "host": target.host,
-                            "port": target.port,
-                            "message": str(exc),
-                        }
-                    )

-            status = "success" if not errors else ("partial" if items else "failed")
-            finalize_rcon_historical_capture_run(
-                run_id,
-                status=status,
-                targets_seen=stats.targets_seen,
-                samples_inserted=stats.samples_inserted,
-                duplicate_samples=stats.duplicate_samples,
-                failed_targets=stats.failed_targets,
-                notes=None if not errors else json.dumps(errors, separators=(",", ":")),
-            )
-        except Exception as exc:
-            finalize_rcon_historical_capture_run(
-                run_id,
-                status="failed",
-                targets_seen=stats.targets_seen,
-                samples_inserted=stats.samples_inserted,
-                duplicate_samples=stats.duplicate_samples,
-                failed_targets=max(1, stats.failed_targets),
-                notes=str(exc),
-            )
-            raise
+def run_rcon_historical_capture_unlocked(
+    *,
+    target_key: str | None = None,
+) -> dict[str, object]:
+    """Capture one prospective RCON sample assuming the shared writer lock is already held."""
+    initialize_rcon_historical_storage()
+    selected_targets = _select_targets(target_key)
+    captured_at = utc_now().isoformat().replace("+00:00", "Z")
+    target_scope = target_key or "all-configured-rcon-targets"
+    run_id = start_rcon_historical_capture_run(mode="capture", target_scope=target_scope)
+    stats = RconHistoricalCaptureStats()
+    items: list[dict[str, object]] = []
+    errors: list[dict[str, object]] = []

-        return {
-            "status": "ok" if items else "error",
-            "run_status": status,
-            "captured_at": captured_at,
-            "target_scope": target_scope,
-            "targets": items,
-            "errors": errors,
-            "storage_status": list_rcon_historical_target_statuses(),
-            "totals": {
-                "targets_seen": stats.targets_seen,
-                "samples_inserted": stats.samples_inserted,
-                "duplicate_samples": stats.duplicate_samples,
-                "failed_targets": stats.failed_targets,
-            },
-        }
+    try:
+        for target in selected_targets:
+            target_metadata = _serialize_target(target)
+            stats.targets_seen += 1
+            try:
+                sample = query_live_server_sample(target)
+                delta = persist_rcon_historical_sample(
+                    run_id=run_id,
+                    captured_at=captured_at,
+                    target=target_metadata,
+                    normalized_payload=sample["normalized"],
+                    raw_payload=sample["raw_session"],
+                )
+                stats.samples_inserted += int(delta["samples_inserted"])
+                stats.duplicate_samples += int(delta["duplicate_samples"])
+                items.append(
+                    {
+                        "target_key": target_metadata["target_key"],
+                        "external_server_id": target.external_server_id,
+                        "captured_at": captured_at,
+                        "sample_inserted": bool(delta["samples_inserted"]),
+                        "normalized": sample["normalized"],
+                    }
+                )
+            except Exception as exc:  # noqa: BLE001 - controlled worker failures
+                stats.failed_targets += 1
+                mark_rcon_historical_capture_failure(
+                    run_id=run_id,
+                    target=target_metadata,
+                    error_message=str(exc),
+                )
+                errors.append(
+                    {
+                        "target_key": target_metadata["target_key"],
+                        "name": target.name,
+                        "host": target.host,
+                        "port": target.port,
+                        "message": str(exc),
+                    }
+                )
+
+        status = "success" if not errors else ("partial" if items else "failed")
+        finalize_rcon_historical_capture_run(
+            run_id,
+            status=status,
+            targets_seen=stats.targets_seen,
+            samples_inserted=stats.samples_inserted,
+            duplicate_samples=stats.duplicate_samples,
+            failed_targets=stats.failed_targets,
+            notes=None if not errors else json.dumps(errors, separators=(",", ":")),
+        )
+    except Exception as exc:
+        finalize_rcon_historical_capture_run(
+            run_id,
+            status="failed",
+            targets_seen=stats.targets_seen,
+            samples_inserted=stats.samples_inserted,
+            duplicate_samples=stats.duplicate_samples,
+            failed_targets=max(1, stats.failed_targets),
+            notes=str(exc),
+        )
+        raise
+
+    return {
+        "status": "ok" if items else "error",
+        "run_status": status,
+        "captured_at": captured_at,
+        "target_scope": target_scope,
+        "targets": items,
+        "errors": errors,
+        "storage_status": list_rcon_historical_target_statuses(),
+        "totals": {
+            "targets_seen": stats.targets_seen,
+            "samples_inserted": stats.samples_inserted,
+            "duplicate_samples": stats.duplicate_samples,
+            "failed_targets": stats.failed_targets,
+        },
+    }


 def run_periodic_rcon_historical_capture(