Hacktoberfest 2026:メンテナが10月に向けて印を付けた、オープンで初心者向けの issue。 Hacktoberfest の issue を見る

[Bug]: One seed that fails the destination check rejects the whole `/crawl` batch

オープン
#2,288 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
3/5
見積もり時間
1〜2日
初心者へのやさしさ
72/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
python
領域
api, backend, security

調査の方向性

Start at _normalize_and_validate_seeds in deploy/docker/api.py around lines 650-657 and trace how /crawl collects results. Then inspect resolve_and_pin in deploy/docker/egress_broker.py around lines 92-96, reproducing the two-seed request from the issue. Done means one refused or unresolvable seed becomes an opaque failed result while other seeds still crawl successfully.

索引モデルが issue の本文から書いたものです。

説明

🐞 Bug 🩺 Needs Triage
crawl4ai version

0.9.4

Expected Behavior

The other seeds are crawled, and the refused one comes back as a failed result carrying the same opaque message, the way a robots.txt refusal already does inside a batch.

Current Behavior

_normalize_and_validate_seeds checks every seed before crawling and raises on the first one refused (api.py#L650-L657), so the whole request fails with 400 {"detail": "URL blocked (SSRF protection): URL blocked"} and no results. resolve_and_pin also raises EgressBlocked when getaddrinfo fails (egress_broker.py#L92-L96), so a host that simply doesn't resolve (a dead domain from a stale sitemap) takes the batch down the same way. The caller can't tell which seed it was, so its only recovery is to split the batch and retry.

Keeping the message opaque, with no resolution oracle, is right, and a per-URL failure keeps it: it says no more than today's 400, because a caller can already find the refused seed by sending the URLs one at a time.

If per-URL failures are the direction you want, I can send a PR. cc @ntohidi

Is this reproducible?

Yes

Inputs Causing the Bug
- urls: ["https://example.com/", "https://no-such-host-12345.example/"]
Steps to Reproduce
1. Run the Docker image (0.9.4, or built from develop), without CRAWL4AI_ALLOW_INTERNAL_URLS
2. POST /crawl with the two URLs above
3. 400, no results; the first URL alone crawls fine
Code snippets
import httpx

r = httpx.post("http://localhost:11235/crawl", headers={"Authorization": f"Bearer {TOKEN}"},
               json={"urls": ["https://example.com/", "https://no-such-host-12345.example/"]}, timeout=60)
print(r.status_code, r.text)   # 400 {"detail":"URL blocked (SSRF protection): URL blocked"}
OS

Linux (Docker image built from develop @ 1f68e5b)

Python version

3.12 (the image's)

Browser

Chromium (Playwright, bundled)

Browser version

No response

Error logs & Screenshots (if applicable)

[seedbatch] one resolvable + one unresolvable seed -> HTTP 400: {"detail":"URL blocked (SSRF protection): URL blocked"}
[seedbatch] the resolvable seed alone -> HTTP 200, success=True

主要言語
Python
スター
84.5k
フォーク
8.7k
平均マージ
3日 9時間
マージ済み PR(30日)
17

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

unclecode/crawl4ai のほかの issue

unclecode/crawl4ai の issue をすべて見る

似ている issue

Python の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。