feat: bound aggregate Gitea transport traffic (#246)
CI / lint (pull_request) Successful in 21s
CI / build-frontend (pull_request) Successful in 5s

This commit is contained in:
timmy committed 2026-08-08 00:34:10 +00:00
1 parent d7d6170c3e
commit 4f6ba0302d
3 files changed
+321 -4

No files matched your search

+119 -4
View File
@@ -2,6 +2,7 @@ import asyncio
import os
import re
import shlex
from contextlib import asynccontextmanager
from typing import Any
from urllib.parse import urlsplit
@@ -12,7 +13,116 @@ GITEA_TOKEN = os.getenv("GITEA_TOKEN", "")
REVIEW_DIFF_MAX_BYTES = 64 * 1024
REVIEW_DIFF_MAX_LINES = 400
AVAILABLE_ISSUE_PAGE_CONCURRENCY = 3
_client: httpx.AsyncClient | None = None
_client: "GiteaTransport | None" = None
class GiteaOverloadedError(RuntimeError):
"""Raised when bounded transport admission expires before capacity is available."""
class GiteaTransport:
"""Application-lifetime HTTP transport with single-flight concurrent reads."""
def __init__(
self,
*,
max_concurrency: int = 8,
admission_timeout: float = 0.25,
**kwargs,
) -> None:
max_concurrency = max(2, max_concurrency)
admission_timeout = max(0.001, admission_timeout)
self._http = httpx.AsyncClient(base_url=GITEA_URL, timeout=10, **kwargs)
self.max_concurrency = max_concurrency
self.admission_timeout = admission_timeout
self._request_slots = asyncio.Semaphore(max(2, max_concurrency))
self._read_slots = asyncio.Semaphore(max(1, max_concurrency - 1))
self._reads: dict[tuple, asyncio.Task[httpx.Response]] = {}
@property
def is_closed(self) -> bool:
return self._http.is_closed
def _read_key(self, url: str, kwargs: dict) -> tuple:
params = tuple(httpx.QueryParams(kwargs.get("params", {})).multi_items())
headers = tuple(sorted(httpx.Headers(kwargs.get("headers", {})).multi_items()))
return url, params, headers
async def _acquire(self, semaphore: asyncio.Semaphore) -> None:
try:
await asyncio.wait_for(semaphore.acquire(), timeout=self.admission_timeout)
except TimeoutError as exc:
raise GiteaOverloadedError("Gitea transport is saturated") from exc
async def _perform_get(self, url: str, kwargs: dict) -> httpx.Response:
await self._acquire(self._read_slots)
try:
await self._acquire(self._request_slots)
try:
return await self._http.get(url, **kwargs)
finally:
self._request_slots.release()
finally:
self._read_slots.release()
async def get(self, url: str, **kwargs) -> httpx.Response:
key = self._read_key(url, kwargs)
task = self._reads.get(key)
if task is None:
task = asyncio.create_task(self._perform_get(url, kwargs))
self._reads[key] = task
task.add_done_callback(
lambda completed, request_key=key: (
self._reads.pop(request_key, None)
if self._reads.get(request_key) is completed else None
)
)
return await asyncio.shield(task)
async def _mutate(self, method: str, url: str, kwargs: dict) -> httpx.Response:
await self._acquire(self._request_slots)
try:
return await self._http.request(method, url, **kwargs)
finally:
self._request_slots.release()
async def post(self, url: str, **kwargs) -> httpx.Response:
return await self._mutate("POST", url, kwargs)
async def patch(self, url: str, **kwargs) -> httpx.Response:
return await self._mutate("PATCH", url, kwargs)
async def put(self, url: str, **kwargs) -> httpx.Response:
return await self._mutate("PUT", url, kwargs)
async def delete(self, url: str, **kwargs) -> httpx.Response:
return await self._mutate("DELETE", url, kwargs)
@asynccontextmanager
async def stream(self, method: str, url: str, **kwargs):
await self._acquire(self._read_slots)
try:
await self._acquire(self._request_slots)
try:
async with self._http.stream(method, url, **kwargs) as response:
yield response
finally:
self._request_slots.release()
finally:
self._read_slots.release()
def __getattr__(self, name: str):
return getattr(self._http, name)
async def aclose(self) -> None:
tasks = list(self._reads.values())
for task in tasks:
if not task.done():
task.cancel()
if tasks:
await asyncio.gather(*tasks, return_exceptions=True)
self._reads.clear()
await self._http.aclose()
class WorkItems(list[dict]):
@@ -58,14 +168,19 @@ def _auth() -> dict[str, str]:
return headers
def start_client(**kwargs) -> httpx.AsyncClient:
def start_client(**kwargs) -> GiteaTransport:
"""Create the application-lifetime Gitea transport."""
global _client
_client = httpx.AsyncClient(base_url=GITEA_URL, timeout=10, **kwargs)
kwargs.setdefault("max_concurrency", int(os.getenv("GITEA_MAX_CONCURRENCY", "8")))
kwargs.setdefault(
"admission_timeout",
float(os.getenv("GITEA_ADMISSION_TIMEOUT_SECONDS", "0.25")),
)
_client = GiteaTransport(**kwargs)
return _client
def _get_client() -> httpx.AsyncClient:
def _get_client() -> GiteaTransport:
if _client is None or _client.is_closed:
return start_client()
return _client