nonebot_bison/metrics.py

@@ -1,8 +1,6 @@
-import time
-
 from fastapi import APIRouter
 from starlette.responses import Response
-from prometheus_client import CONTENT_TYPE_LATEST, Gauge, Counter, Histogram, generate_latest
+from prometheus_client import CONTENT_TYPE_LATEST, Counter, generate_latest
 
 # Request counter
 request_counter = Counter(
@@ -12,28 +10,6 @@ request_counter = Counter(
 # Sent counter
 sent_counter = Counter("bison_sent_counter", "The number of sent messages", ["site_name", "platform_name", "target"])
 
-cookie_choose_counter = Counter(
-    "bison_cookie_choose_counter", "The number of cookie choose", ["site_name", "target", "cookie_id"]
-)
-
-request_histogram = Histogram(
-    "bison_request_histogram",
-    "The time of platform used to request the source",
-    ["site_name", "platform_name"],
-    buckets=[0.1, 0.5, 1, 2, 5, 10, 30, 60],
-)
-
-render_histogram = Histogram(
-    "bison_render_histogram",
-    "The time of theme used to render",
-    ["site_name", "platform_name"],
-    buckets=[0.1, 0.5, 1, 2, 5, 10, 30, 60],
-)
-
-start_time = Gauge("bison_start_time", "The start time of the program")
-start_time.set(time.time())
-
-
 metrics_router = APIRouter(prefix="/api/metrics", tags=["metrics"])
 
 

nonebot_bison/scheduler/scheduler.py

@@ -1,14 +1,12 @@
 from dataclasses import dataclass
 from collections import defaultdict
-from collections.abc import Callable
 
 from nonebot.log import logger
 from nonebot_plugin_apscheduler import scheduler
-from apscheduler.events import EVENT_JOB_MAX_INSTANCES
 from nonebot_plugin_saa.utils.exceptions import NoBotFound
 
 from nonebot_bison.utils import ClientManager
-from nonebot_bison.metrics import sent_counter, request_counter, render_histogram, request_histogram
+from nonebot_bison.metrics import sent_counter, request_counter
 
 from ..config import config
 from ..send import send_msgs
@@ -26,15 +24,6 @@ class Schedulable:
     use_batch: bool = False
 
 
-def handle_time_exceeded(event):
-    # event.job_id 是该任务在 apscheduler 的 id, 进而可以获得该任务的函数，再获取该函数绑定的对象
-    logger.warning(f"{scheduler.get_job(event.job_id).func.__self__.name} 抓取执行超时")
-    scheduler.get_job(event.job_id).func.__self__.metrics_report(False)
-
-
-scheduler.add_listener(handle_time_exceeded, EVENT_JOB_MAX_INSTANCES)
-
-
 class Scheduler:
     schedulable_list: list[Schedulable]  # for load weigth from db
     batch_api_target_cache: dict[str, dict[Target, list[Target]]]  # platform_name -> (target -> [target])
@@ -67,7 +56,6 @@ class Scheduler:
 
         self.platform_name_list = platform_name_list
         self.pre_weight_val = 0  # 轮调度中“本轮”增加权重和的初值
-        self.metrics_report: Callable[[bool], None] | None = None  # 作为函数变量，允许外部调用来上报此次抓取是否成功
         logger.info(
             f"register scheduler for {self.name} with "
             f"{self.scheduler_config.schedule_type} {self.scheduler_config.schedule_setting}"
@@ -109,30 +97,20 @@ class Scheduler:
 
         success_flag = False
         platform_obj = platform_manager[schedulable.platform_name](context)
-        # 通过闭包的形式，将此次抓取任务的信息保存为函数变量，允许在该任务无法正常结束时由外部上报
-        self.metrics_report = lambda x: request_counter.labels(
-            platform_name=schedulable.platform_name,
-            site_name=platform_obj.site.name,
-            target=schedulable.target,
-            success=x,
-        ).inc()
         try:
-            with request_histogram.labels(
+            if schedulable.use_batch:
-                platform_name=schedulable.platform_name, site_name=platform_obj.site.name
+                batch_targets = self.batch_api_target_cache[schedulable.platform_name][schedulable.target]
-            ).time():
+                sub_units = []
-                if schedulable.use_batch:
+                for batch_target in batch_targets:
-                    batch_targets = self.batch_api_target_cache[schedulable.platform_name][schedulable.target]
+                    userinfo = await config.get_platform_target_subscribers(schedulable.platform_name, batch_target)
-                    sub_units = []
+                    sub_units.append(SubUnit(batch_target, userinfo))
-                    for batch_target in batch_targets:
+                to_send = await platform_obj.do_batch_fetch_new_post(sub_units)
-                        userinfo = await config.get_platform_target_subscribers(schedulable.platform_name, batch_target)
+            else:
-                        sub_units.append(SubUnit(batch_target, userinfo))
+                send_userinfo_list = await config.get_platform_target_subscribers(
-                    to_send = await platform_obj.do_batch_fetch_new_post(sub_units)
+                    schedulable.platform_name, schedulable.target
-                else:
+                )
-                    send_userinfo_list = await config.get_platform_target_subscribers(
+                to_send = await platform_obj.do_fetch_new_post(SubUnit(schedulable.target, send_userinfo_list))
-                        schedulable.platform_name, schedulable.target
+            success_flag = True
-                    )
-                    to_send = await platform_obj.do_fetch_new_post(SubUnit(schedulable.target, send_userinfo_list))
-                success_flag = True
         except SkipRequestException as err:
             logger.debug(f"skip request: {err}")
         except Exception as err:
@@ -142,23 +120,27 @@ class Scheduler:
             err.args += (records,)
             raise
 
-        self.metrics_report(success_flag)
+        request_counter.labels(
+            platform_name=schedulable.platform_name,
+            site_name=platform_obj.site.name,
+            target=schedulable.target,
+            success=success_flag,
+        ).inc()
         if not to_send:
             return
         sent_counter.labels(
             platform_name=schedulable.platform_name, site_name=platform_obj.site.name, target=schedulable.target
         ).inc()
-        with render_histogram.labels(platform_name=schedulable.platform_name, site_name=platform_obj.site.name).time():
+        for user, send_list in to_send:
-            for user, send_list in to_send:
+            for send_post in send_list:
-                for send_post in send_list:
+                logger.info(f"send to {user}: {send_post}")
-                    logger.info(f"send to {user}: {send_post}")
+                try:
-                    try:
+                    await send_msgs(
-                        await send_msgs(
+                        user,
-                            user,
+                        await send_post.generate_messages(),
-                            await send_post.generate_messages(),
+                    )
-                        )
+                except NoBotFound:
-                    except NoBotFound:
+                    logger.warning("no bot connected")
-                        logger.warning("no bot connected")
 
     def insert_new_schedulable(self, platform_name: str, target: Target):
         self.pre_weight_val += 1000

nonebot_bison/utils/site.py

@@ -13,7 +13,6 @@ from ..types import Target
 from ..config import config
 from .http import http_client
 from ..config.db_model import Cookie
-from ..metrics import cookie_choose_counter
 
 
 class ClientManager(ABC):
@@ -132,7 +131,6 @@ class CookieClientManager(ClientManager):
         """获取 client，根据 target 选择 cookie"""
         client = http_client()
         cookie = await self._choose_cookie(target)
-        cookie_choose_counter.labels(site_name=self._site_name, target=target, cookie_id=cookie.id).inc()
         if cookie.is_universal:
             logger.trace(f"平台 {self._site_name} 未获取到用户cookie, 使用匿名cookie")
         else: