feat(监控): 完善系统监控、登录状态与访问审计能力
Client Quality Gates / Shared client and Web (push) Has been cancelled
Client Quality Gates / macOS Desktop (push) Has been cancelled

This commit is contained in:
Cheng Zhou
2026-07-10 17:11:24 +08:00
parent 400c9be3a7
commit f11a5c84d9
82 changed files with 10283 additions and 1781 deletions
+59 -3
View File
@@ -5,8 +5,9 @@ import time
from contextlib import asynccontextmanager
from collections import defaultdict
from fastapi import FastAPI
from fastapi import Depends, FastAPI
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse
from sqlalchemy import text
from app.api.v1.router import api_router
@@ -19,15 +20,23 @@ from app.db.session import SessionLocal, engine
from app.services.visit_scheduler import run_daily_lost_visit_job
from app.services.permission_log_writer import start_log_writer, stop_log_writer
from app.services.permission_metric_aggregator import run_hourly_metric_aggregation
from app.services.source_location_aggregator import run_hourly_source_location_aggregation
from app.services.monitoring_server_location import resolve_monitoring_server_location
from app.services.monitoring_retention import run_monitoring_retention
from app.services.security_access_log_writer import (
get_security_log_writer,
start_security_log_writer,
stop_security_log_writer,
)
from app.core.security import decode_token
from app.core.deps import get_db_session
from app.core.request_context import (
build_request_audit_context,
build_request_snapshot,
build_sanitized_request_headers,
get_request_audit_context,
reset_request_audit_context,
resolve_ctms_client_type,
resolve_client_ip,
set_request_audit_context,
)
@@ -42,8 +51,12 @@ setup_config_stats: dict[str, int] = defaultdict(int)
@asynccontextmanager
async def lifespan(_: FastAPI):
stop_event = asyncio.Event()
await resolve_monitoring_server_location()
scheduler_task = asyncio.create_task(run_daily_lost_visit_job(stop_event))
aggregator_task = asyncio.create_task(run_hourly_metric_aggregation(stop_event))
source_location_aggregator_task = asyncio.create_task(
run_hourly_source_location_aggregation(stop_event)
)
await start_log_writer()
await start_security_log_writer()
if settings.ENV == "development":
@@ -52,12 +65,15 @@ async def lifespan(_: FastAPI):
await conn.run_sync(Base.metadata.create_all)
async with SessionLocal() as session:
await ensure_admin_exists(session)
retention_task = asyncio.create_task(run_monitoring_retention(stop_event))
yield
stop_event.set()
await stop_log_writer()
await stop_security_log_writer()
await scheduler_task
await aggregator_task
await source_location_aggregator_task
await retention_task
async def _ensure_legacy_primary_keys(conn) -> None:
@@ -129,19 +145,23 @@ def create_app() -> FastAPI:
"Accept",
"Authorization",
"Content-Type",
"X-CTMS-Client-Source",
"X-CTMS-Client-Type",
"X-CTMS-Client-Version",
"X-CTMS-Client-Platform",
"X-CTMS-Build-Channel",
"X-CTMS-Build-Commit",
"X-Request-ID",
"X-Correlation-ID",
],
expose_headers=["X-Request-ID"],
)
@app.middleware("http")
async def setup_config_monitoring_middleware(request, call_next):
path = request.url.path
is_setup_config_path = "/api/v1/studies/" in path and "/setup-config" in path
should_security_log = path.startswith("/api/")
should_security_log = path.startswith("/api/") and path != "/api/v1/auth/session/heartbeat"
started_at = time.perf_counter()
audit_context = build_request_audit_context(request)
audit_context_token = set_request_audit_context(audit_context)
@@ -167,6 +187,8 @@ def create_app() -> FastAPI:
raise
status_code = int(response.status_code)
if audit_context.request_id:
response.headers["X-Request-ID"] = audit_context.request_id
if is_setup_config_path:
normalized_path = UUID_RE.sub("{study_id}", path)
duration_ms = int((time.perf_counter() - started_at) * 1000)
@@ -229,6 +251,36 @@ def create_app() -> FastAPI:
async def health() -> dict[str, str]:
return {"status": "ok"}
@app.get(
"/readyz",
tags=["health"],
summary="服务就绪检查",
description="验证应用可访问数据库;失败时返回 503。",
)
async def readiness(db=Depends(get_db_session)):
started_at = time.perf_counter()
try:
await db.execute(text("SELECT 1"))
except Exception as exc:
logger.exception("Readiness database check failed")
return JSONResponse(
status_code=503,
content={
"status": "not_ready",
"database": {
"status": "unhealthy",
"error_type": type(exc).__name__,
},
},
)
return {
"status": "ready",
"database": {
"status": "healthy",
"latency_ms": round((time.perf_counter() - started_at) * 1000, 2),
},
}
@app.get(
"/health/setup-config-stats",
tags=["health"],
@@ -271,6 +323,7 @@ def _enqueue_security_access_log(request, path: str, status_code: int, started_a
if not writer:
return
auth_status, user_identifier = _resolve_auth_context(request)
context = get_request_audit_context()
writer.enqueue(
{
"method": request.method,
@@ -279,11 +332,14 @@ def _enqueue_security_access_log(request, path: str, status_code: int, started_a
"elapsed_ms": round((time.perf_counter() - started_at) * 1000, 2),
"client_ip": resolve_client_ip(request),
"user_agent": request.headers.get("user-agent"),
"client_type": request.headers.get("x-ctms-client-type"),
"client_type": resolve_ctms_client_type(request.headers),
"client_version": request.headers.get("x-ctms-client-version"),
"client_platform": request.headers.get("x-ctms-client-platform"),
"build_channel": request.headers.get("x-ctms-build-channel"),
"build_commit": request.headers.get("x-ctms-build-commit"),
"request_headers": context.request_headers if context else build_sanitized_request_headers(request),
"request_snapshot": context.request_snapshot if context else build_request_snapshot(request),
"request_id": context.request_id if context else None,
"auth_status": auth_status,
"user_identifier": user_identifier,
}