feat: record the reasoning verdict in telemetry

This issue surfaced only because someone ran a slow suite that is
excluded by default and had not been run for eighteen days. As a column
it becomes a query: which model stopped being observable, and when.

The emitter unwraps the enum to a plain str at the single _record exit.
asyncpg makes no promise about encoding a str subclass, and a telemetry
write that fails is downgraded to one warning — it would not crash, it
would just quietly cost the Postgres path a column. Normalising at the
emitter follows what tenant_id, meta and sampling already do.

The column is appended last in COLUMNS and in both DDLs. An existing
table can only take ALTER at the end, so putting it anywhere else
forks the physical column order between a freshly built database and a
backfilled one.

src/polygateway/telemetry/schema.py

@@ -5,8 +5,8 @@
 多处各存一份必然漂移,而漂移的表现是"下游照打印的 SQL 建完表,库仍报缺列"。
 
 **`COLUMNS` 是 INSERT 字段序,不是物理列序**: 数据库自填的 `created_at` 不在其中(它带
-`DEFAULT now()` / `datetime('now')`,库从不显式写它)。物理表列 = 24 个 INSERT 字段 +
-`created_at` = 25;列数断言一律按物理列数写,两套口径混用是最易错处。
+`DEFAULT now()` / `datetime('now')`,库从不显式写它)。物理表列 = 25 个 INSERT 字段 +
+`created_at` = 26;列数断言一律按物理列数写,两套口径混用是最易错处。
 
 本模块只依赖标准库: `telemetry/` 与 `backends/`、`transports/`、`structured/` 同层且
 互不依赖(import-linter 契约执法)。
@@ -50,7 +50,8 @@ CREATE TABLE IF NOT EXISTS llm_calls (
     sampling             TEXT,
     reasoning_tokens     INTEGER,
     tenant_id            TEXT    NOT NULL DEFAULT '',
-    meta                 TEXT    NOT NULL DEFAULT '{}'
+    meta                 TEXT    NOT NULL DEFAULT '{}',
+    thinking_observation TEXT
 );
 """
 
@@ -80,7 +81,8 @@ CREATE TABLE IF NOT EXISTS llm_calls (
     sampling             TEXT,
     reasoning_tokens     INTEGER,
     tenant_id            TEXT NOT NULL DEFAULT '',
-    meta                 JSONB NOT NULL DEFAULT '{}'::jsonb
+    meta                 JSONB NOT NULL DEFAULT '{}'::jsonb,
+    thinking_observation TEXT
 );
 """
 
@@ -95,6 +97,9 @@ SQLITE_BACKFILL = (
     # ("Cannot add a NOT NULL column with default value NULL"),补列全盘失败。
     ("tenant_id", "TEXT NOT NULL DEFAULT ''"),
     ("meta", "TEXT NOT NULL DEFAULT '{}'"),
+    # 可空: 补列之前的行没有裁定结果,NULL 如实表达"这行根本没记过这件事",
+    # 与哨兵串 'unknown'(库确实裁过但判不出来)是两回事,不得混同
+    ("thinking_observation", "TEXT"),
 )
 
 # PG 补列的列定义。语句由此派生成两份文本(见下),使"库内执行的那份"与"打印给
@@ -107,6 +112,8 @@ _PG_BACKFILL_DECLS = (
     # 两个默认值都是非易失常量,PG 11+ 只改 catalog 不重写全表,故大表补列亦是秒级
     ("tenant_id", "TEXT NOT NULL DEFAULT ''"),
     ("meta", "JSONB NOT NULL DEFAULT '{}'::jsonb"),
+    # 可空,理由同 SQLITE_BACKFILL 同名项
+    ("thinking_observation", "TEXT"),
 )
 
 # 新列排在 created_at 之后: 与旧表 ALTER 追加的位置一致(见 SQLITE_BACKFILL 同款注释)。
@@ -143,6 +150,7 @@ COLUMNS = (
     "reasoning_tokens",
     "tenant_id",
     "meta",
+    "thinking_observation",
 )
 
 _COLUMN_SET = frozenset(COLUMNS)

src/polygateway/telemetry/sqlite.py

@@ -143,7 +143,7 @@ class SQLiteRecorder:
                     logger.warning("SQLite 遥测补列失败(写入将逐行降级): {}", exc)
 
     async def record_llm_call(self, **fields: object) -> None:
-        """写一行遥测;字段集合即 24 字段冻结签名(ports.TelemetryRecorder)。
+        """写一行遥测;字段集合即 25 字段冻结签名(ports.TelemetryRecorder)。
 
         取值按 `self._columns`(manual 档可能已被裁剪),与 `self._insert` 的
         占位符同序——两者必须一起改,分开改就是把值写进错位的列。