feat: add postgres telemetry recorder with two-tier degradation

This commit is contained in:
2026-07-21 00:50:33 -04:00
parent 0e22fcf433
commit abb65c2324
5 changed files with 359 additions and 12 deletions
+40 -8
View File
@@ -162,6 +162,10 @@ class GatewayClient:
transport_aclose = getattr(self._transport, "aclose", None) transport_aclose = getattr(self._transport, "aclose", None)
if transport_aclose is not None: if transport_aclose is not None:
await transport_aclose() await transport_aclose()
telemetry_aclose = getattr(self._telemetry, "aclose", None)
if telemetry_aclose is not None:
await telemetry_aclose() # Postgres 等异步后端
else:
telemetry_close = getattr(self._telemetry, "close", None) telemetry_close = getattr(self._telemetry, "close", None)
if telemetry_close is not None: if telemetry_close is not None:
telemetry_close() telemetry_close()
@@ -196,14 +200,8 @@ class GatewayClient:
scope=settings.scope, scope=settings.scope,
sources=sources, sources=sources,
selector=_build_selector(settings.selector), selector=_build_selector(settings.selector),
limiter=limiter limiter=limiter or _build_limiter(settings, sources),
or InMemoryLimiter( breaker=breaker or _build_breaker(settings),
scope=settings.scope,
sources={s.name: s for s in sources},
global_limits=settings.global_limits,
lease_ttl_s=settings.lease_ttl_s,
),
breaker=breaker or InMemoryGate(config=settings.breaker),
transport=OpenAICompatTransport(registry=registry), transport=OpenAICompatTransport(registry=registry),
retry=settings.retry, retry=settings.retry,
backpressure=settings.backpressure, backpressure=settings.backpressure,
@@ -240,6 +238,35 @@ class GatewayClient:
) )
def _build_limiter(settings: GatewaySettings, sources: list[SourceConfig]) -> RateLimiter:
if settings.limiter_backend == "redis":
from polygateway.backends.redis.limiter import RedisLimiter
assert settings.redis_url is not None # 内部不变量: config 已校验
return RedisLimiter.from_url(
settings.redis_url,
scope=settings.scope,
sources={s.name: s for s in sources},
global_limits=settings.global_limits,
lease_ttl_s=settings.lease_ttl_s,
)
return InMemoryLimiter(
scope=settings.scope,
sources={s.name: s for s in sources},
global_limits=settings.global_limits,
lease_ttl_s=settings.lease_ttl_s,
)
def _build_breaker(settings: GatewaySettings) -> ProviderGate:
if settings.breaker_backend == "redis":
from polygateway.backends.redis.breaker import RedisGate
assert settings.redis_url is not None # 内部不变量: config 已校验
return RedisGate.from_url(settings.redis_url, config=settings.breaker, scope=settings.scope)
return InMemoryGate(config=settings.breaker)
def _build_selector(name: str) -> SourceSelector: def _build_selector(name: str) -> SourceSelector:
return RoundRobinSelector() if name == "round_robin" else LeastInflightSelector() return RoundRobinSelector() if name == "round_robin" else LeastInflightSelector()
@@ -258,6 +285,11 @@ def _build_cache(settings: GatewaySettings) -> CacheBackend | None:
def _build_telemetry(settings: GatewaySettings) -> TelemetryRecorder | None: def _build_telemetry(settings: GatewaySettings) -> TelemetryRecorder | None:
if settings.telemetry_backend == "none": if settings.telemetry_backend == "none":
return None return None
if settings.telemetry_backend == "postgres":
from polygateway.telemetry.postgres import PostgresRecorder
assert settings.telemetry_pg_dsn is not None # 内部不变量: config 已校验
return PostgresRecorder(settings.telemetry_pg_dsn)
from polygateway.telemetry.sqlite import SQLiteRecorder from polygateway.telemetry.sqlite import SQLiteRecorder
assert settings.telemetry_sqlite_path is not None # 内部不变量: config 已校验 assert settings.telemetry_sqlite_path is not None # 内部不变量: config 已校验
+136
View File
@@ -0,0 +1,136 @@
"""Postgres 遥测后端(M2 设计 §5): asyncpg lazy 池 + 两级降级。
参考仓无先例(三项目遥测全 SQLite);asyncpg 工程写法取 GovDoc
`taskrun/postgres_store.py`($n 占位、`CREATE TABLE IF NOT EXISTS`、
`ON CONFLICT DO NOTHING`),但其"失败冒泡"方向按遥测铁律**有意反转**:
① 结构性失败(建池/建表)→ warning 一次后永久降级(池置 None 短路);
② 运行时单条写失败 → 逐条 warning 丢弃,不降级不重试(连接抖动由
asyncpg 池自恢复;避免浸泡开头一次抖动导致后续全程失遥测)。
构造不连库(lazy),18 列 schema 与 SQLite 版同名同序。
"""
from __future__ import annotations
import asyncio
from typing import TYPE_CHECKING
from loguru import logger
if TYPE_CHECKING:
import asyncpg
_DDL = """
CREATE TABLE IF NOT EXISTS llm_calls (
call_id TEXT PRIMARY KEY,
parent_call_id TEXT,
session_id TEXT,
model TEXT NOT NULL,
provider TEXT NOT NULL,
source_name TEXT NOT NULL,
messages TEXT NOT NULL,
response TEXT NOT NULL,
thinking TEXT NOT NULL DEFAULT '',
prompt_tokens INTEGER NOT NULL,
completion_tokens INTEGER NOT NULL,
usage_source TEXT NOT NULL,
latency_ms INTEGER NOT NULL,
ttft_ms DOUBLE PRECISION,
max_inter_token_ms DOUBLE PRECISION,
cache_hit BOOLEAN NOT NULL DEFAULT FALSE,
error TEXT,
cost DOUBLE PRECISION,
created_at TIMESTAMPTZ NOT NULL DEFAULT now()
);
"""
_COLUMNS = (
"call_id",
"parent_call_id",
"session_id",
"model",
"provider",
"source_name",
"messages",
"response",
"thinking",
"prompt_tokens",
"completion_tokens",
"usage_source",
"latency_ms",
"ttft_ms",
"max_inter_token_ms",
"cache_hit",
"error",
"cost",
)
_INSERT = (
f"INSERT INTO llm_calls ({', '.join(_COLUMNS)}) "
f"VALUES ({', '.join(f'${i + 1}' for i in range(len(_COLUMNS)))}) "
"ON CONFLICT (call_id) DO NOTHING"
)
class PostgresRecorder:
"""TelemetryRecorder 端口的 Postgres 实现;asyncpg 原生异步,无线程桥接。"""
def __init__(self, dsn: str, *, pool: asyncpg.Pool | None = None) -> None:
try:
import asyncpg # noqa: F401 - 仅探测 extra 是否安装
except ImportError as exc:
raise ImportError(
"Postgres 遥测未启用: 安装 pip install 'polygateway[postgres]' 后重试"
) from exc
self._dsn = dsn
self._pool: asyncpg.Pool | None = pool
self._external_pool = pool is not None
self._schema_ready = False
self._failed = False # 结构性降级标志: 置位后所有写入短路
self._init_lock = asyncio.Lock()
async def _ensure_ready(self) -> asyncpg.Pool | None:
"""lazy 建池+建表;结构性失败 warning 一次后永久降级(设计 §5 两级之一)。"""
if self._failed:
return None
if self._schema_ready:
return self._pool
async with self._init_lock:
if self._failed or self._schema_ready:
return None if self._failed else self._pool
try:
if self._pool is None:
import asyncpg
self._pool = await asyncpg.create_pool(self._dsn, timeout=10)
async with self._pool.acquire() as conn:
await conn.execute(_DDL)
self._schema_ready = True
return self._pool
except asyncio.CancelledError:
raise
except Exception as exc:
self._failed = True
logger.warning("Postgres 遥测初始化失败,后续记录降级为 no-op: {}", exc)
return None
async def record_llm_call(self, **fields: object) -> None:
"""写一行遥测;单条失败逐条 warning 丢弃(两级降级之二),绝不冒泡。"""
pool = await self._ensure_ready()
if pool is None:
return
row = tuple(fields[col] for col in _COLUMNS)
try:
async with pool.acquire() as conn:
await conn.execute(_INSERT, *row)
except asyncio.CancelledError:
raise
except Exception as exc:
# 遥测铁律: 丢一条 < 拖垮调用;仅记 warning(非 pass),池自恢复
logger.warning("Postgres 遥测写入失败(丢弃该行): {}", exc)
async def aclose(self) -> None:
"""幂等关闭自建池;注入的池归注入方管理。"""
pool, self._pool = self._pool, None
self._schema_ready = False
if pool is not None and not self._external_pool:
await pool.close()
@@ -0,0 +1,159 @@
"""PostgresRecorder 集成测试(M2 设计 §5;真实实验室 Postgres,polygateway 专用库)。
DSN 走 .env `PGW_TELEMETRY_PG_DSN`,缺则 skip。该实例上有 app/chs_prod 等
在用库——本测试只允许连 polygateway 专用库(fixture 里守卫)。
"""
from __future__ import annotations
import asyncio
import os
import pytest
from dotenv import dotenv_values
from polygateway.telemetry.postgres import PostgresRecorder
_EXPECTED_COLUMNS = [
"call_id",
"parent_call_id",
"session_id",
"model",
"provider",
"source_name",
"messages",
"response",
"thinking",
"prompt_tokens",
"completion_tokens",
"usage_source",
"latency_ms",
"ttft_ms",
"max_inter_token_ms",
"cache_hit",
"error",
"cost",
"created_at",
]
def _dsn() -> str | None:
merged = {**dotenv_values(".env"), **os.environ}
raw = merged.get("PGW_TELEMETRY_PG_DSN")
if not raw:
return None
scheme, sep, rest = raw.partition("://")
return f"{scheme.partition('+')[0]}{sep}{rest}"
@pytest.fixture
async def dsn():
value = _dsn()
if value is None:
pytest.skip("PGW_TELEMETRY_PG_DSN 未配置")
# 隔离守卫: 该实例有 app/chs_prod/mimiciv 等在用库,只许打 polygateway 专用库
if not value.rstrip("/").endswith("/polygateway"):
pytest.fail(f"遥测测试只允许连 polygateway 专用库,当前 DSN 库名不符: {value!r}")
import asyncpg
conn = await asyncpg.connect(value, timeout=10)
try:
await conn.execute("DROP TABLE IF EXISTS llm_calls")
finally:
await conn.close()
return value
async def _record_minimal(recorder: PostgresRecorder, call_id: str = "c1", **overrides) -> None:
fields = {
"call_id": call_id,
"parent_call_id": None,
"session_id": "sess-1",
"model": "m",
"provider": "p",
"source_name": "s1",
"messages": "[]",
"response": "ok",
"thinking": "",
"prompt_tokens": 1,
"completion_tokens": 2,
"usage_source": "measured",
"latency_ms": 10,
"ttft_ms": None,
"max_inter_token_ms": None,
"cache_hit": False,
"error": None,
"cost": None,
}
fields.update(overrides)
await recorder.record_llm_call(**fields)
async def _fetch(dsn: str, sql: str):
import asyncpg
conn = await asyncpg.connect(dsn, timeout=10)
try:
return await conn.fetch(sql)
finally:
await conn.close()
class TestSchema:
async def test_schema_has_frozen_columns_in_order(self, dsn):
recorder = PostgresRecorder(dsn)
try:
await _record_minimal(recorder)
rows = await _fetch(
dsn,
"SELECT column_name FROM information_schema.columns "
"WHERE table_name='llm_calls' ORDER BY ordinal_position",
)
assert [r["column_name"] for r in rows] == _EXPECTED_COLUMNS
finally:
await recorder.aclose()
async def test_call_id_idempotent(self, dsn):
recorder = PostgresRecorder(dsn)
try:
await _record_minimal(recorder, call_id="dup")
await _record_minimal(recorder, call_id="dup", response="second")
rows = await _fetch(dsn, "SELECT response FROM llm_calls WHERE call_id='dup'")
assert [r["response"] for r in rows] == ["ok"] # ON CONFLICT DO NOTHING
finally:
await recorder.aclose()
async def test_concurrent_writes_all_land(self, dsn):
recorder = PostgresRecorder(dsn)
try:
await asyncio.gather(*(_record_minimal(recorder, call_id=f"c{i}") for i in range(50)))
rows = await _fetch(dsn, "SELECT count(*) AS n FROM llm_calls")
assert rows[0]["n"] == 50
finally:
await recorder.aclose()
class TestDegradation:
async def test_unreachable_server_degrades_silently(self):
"""结构性失败(建池不通)→ warning 一次后永久降级,业务零感知。"""
recorder = PostgresRecorder("postgresql://u:p@127.0.0.1:1/x")
await _record_minimal(recorder) # 不抛
await _record_minimal(recorder, call_id="c2") # 已降级短路,同样不抛
await recorder.aclose()
async def test_row_failure_does_not_poison_later_rows(self, dsn):
"""运行时单条写失败(NUL 字节文本被 PG 拒)→ 丢该行,后续行照常落库。"""
recorder = PostgresRecorder(dsn)
try:
await _record_minimal(recorder, call_id="bad", response="nul\x00byte")
await _record_minimal(recorder, call_id="good")
rows = await _fetch(dsn, "SELECT call_id FROM llm_calls ORDER BY call_id")
assert [r["call_id"] for r in rows] == ["good"]
finally:
await recorder.aclose()
async def test_aclose_idempotent(self, dsn):
recorder = PostgresRecorder(dsn)
await _record_minimal(recorder)
await recorder.aclose()
await recorder.aclose()
+16
View File
@@ -130,6 +130,22 @@ class TestFactories:
with pytest.raises(ValueError, match="glm"): with pytest.raises(ValueError, match="glm"):
GatewayClient.from_env("LLM", env=env) GatewayClient.from_env("LLM", env=env)
def test_from_env_builds_redis_governance_backends(self):
"""M2: 配置取 redis 时装配出 Redis 后端(构造不连库,unit 可测)。"""
from polygateway.backends.redis.breaker import RedisGate
from polygateway.backends.redis.limiter import RedisLimiter
from polygateway.client import _build_breaker, _build_limiter
env = dict(
_ENV,
PGW_LIMITER_BACKEND="redis",
PGW_BREAKER_BACKEND="redis",
REDIS_URL="redis://:pw@10.0.0.1:6379/3",
)
settings = GatewaySettings.from_env("LLM", env=env)
assert isinstance(_build_limiter(settings, list(settings.sources)), RedisLimiter)
assert isinstance(_build_breaker(settings), RedisGate)
def test_from_settings_respects_injection(self): def test_from_settings_respects_injection(self):
settings = GatewaySettings.from_env("LLM", env=_ENV) settings = GatewaySettings.from_env("LLM", env=_ENV)
shared = InMemoryLimiter( shared = InMemoryLimiter(
+5 -1
View File
@@ -266,5 +266,9 @@ def test_single_emitter_discipline():
text=True, text=True,
cwd=Path(__file__).resolve().parents[2], cwd=Path(__file__).resolve().parents[2],
).stdout.splitlines() ).stdout.splitlines()
callers = [p for p in out if not p.endswith(("ports.py", "telemetry/sqlite.py"))] callers = [
p
for p in out
if not p.endswith(("ports.py", "telemetry/sqlite.py", "telemetry/postgres.py"))
]
assert callers == ["src/polygateway/middleware/telemetry.py"] assert callers == ["src/polygateway/middleware/telemetry.py"]