feat(sites): 站点请求失败留样(site_failure_sample,每站最近 5 份)

- 迁移 0077 新表 site_failure_sample:类型(parse_error / challenge)、请求类型、脱敏 URL、状态码、
  关键响应头、错误文案、命中规则、脱敏页面开头(≤ 64 KB);样本几十 KB,不放进每条请求都读的 site.config
- 适配器 onFailureSample 回调接到 runtime.siteAdapter,异步写入并删到最近 5 份,写失败只记日志
- GET /sites/{id}/failure-samples(不带正文)与 /failure-samples/{sampleId}(带正文)

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Cxwgnn2pbzQ7rXv3rq7fei
This commit is contained in:
rootandClaude Opus 5.5 committed 2026-10-06 15:41:00 +08:00
1 parent 1779e1a0f2
commit 0088c1c9e7
10 files changed
+355

No files matched your search

@@ -17,6 +17,8 @@ import {
SiteCredentialsPut,
SiteCredentialsResult,
SiteDefinitionView,
SiteFailureSample,
SiteFailureSampleSummary,
SiteHealthResult,
SiteRefreshLoginAccepted,
SiteRefreshLoginInput,
@@ -47,6 +49,7 @@ import {
userDataHistory,
} from "../service/account.ts";
import { audit, closeDecisions, type Exec } from "../service/common.ts";
import { getFailureSample, listFailureSamples } from "../service/failure-sample.ts";
import { feedDefaultsForSite } from "../service/feeds.ts";
import { assignReleaseIdentity } from "../service/identity-admin.ts";
import { ensureRssSite, requireCapability, siteToApi, uniqueSiteKey } from "../site-kinds.ts";
@@ -376,6 +379,37 @@ export function sourcesRoutes(app: ModuleApp, rt: () => SourcesRuntime) {
},
);
route(
app,
{
method: "get",
path: "/sites/{id}/failure-samples",
scope: "read",
summary: "站点请求失败留样:最近几次页面结构无法识别 / 人机验证时的脱敏页面(列表不带正文)",
tags: T_SITES,
params: IdParam,
response: z.object({ items: z.array(SiteFailureSampleSummary) }),
},
async ({ params }) => ({ items: await listFailureSamples(rt().ctx, params.id) }),
);
route(
app,
{
method: "get",
path: "/sites/{id}/failure-samples/{sampleId}",
scope: "read",
summary: "一份失败留样(含脱敏后的页面正文)",
tags: T_SITES,
params: z.object({ id: Id, sampleId: Id }),
response: SiteFailureSample,
},
async ({ params }) => {
const s = await getFailureSample(rt().ctx, params.id, params.sampleId);
if (!s) throw new ApiException("not_found", "样本不存在");
return s;
},
);
// ── W9-A 站点中心:删除、健康检测、刷新登录、用户数据、凭证 ──
route(
app,
@@ -26,6 +26,7 @@ import {
} from "../../shared-cache.ts";
import type { ServerContext } from "../types.ts";
import { CloudflareGate, withClearance } from "./service/cloudflare.ts";
import { failureSampleSink } from "./service/failure-sample.ts";
import { readBrowserUserAgent, resolveUserAgent, type UserAgentDefaults } from "./service/user-agent.ts";
// 模块运行时依赖:站点适配器与限流器(每站一个实例,令牌在 PG 跨进程共享)、RSS 抓取器、TMDB 客户端、时钟。
@@ -333,6 +334,8 @@ export class SourcesRuntime {
// 每条请求都经 siteFetch:代理 + 按账号现场取 UA(BR-4)
fetch: this.siteFetch({ id: site.id, proxy }, defaults, () => this.siteUserAgent(site.id)),
...(baseUrl ? { baseUrl } : {}),
// parse_error / challenge 的脱敏页面样本存进 site_failure_sample(每站最近 5 份)
onFailureSample: failureSampleSink(this.ctx, site.id),
});
this.adapters.set(site.id, { sig, limiter, adapter });
return adapter;
@@ -0,0 +1,83 @@
import type { SiteFailureSample as AdapterSample } from "@umt/adapters";
import type { SiteFailureSample, SiteFailureSampleSummary } from "@umt/contracts";
import { sql } from "@umt/db";
import type { ServerContext } from "../../types.ts";
import { toApi } from "../rows.ts";
// 站点请求失败留样(表 site_failure_sample,迁移 0077):适配器报 parse_error / challenge 时交出脱敏页面样本
// (SiteAdapterDeps.onFailureSample),这里存下来,每站只留最近 SAMPLES_PER_SITE 份。站点页「技术信息」里看。
// 样本在适配器里已脱敏(个人信息块、passkey / authkey / uid / Cookie / 邮箱),这里不再处理正文。
/** 每站保留的样本数 */
export const SAMPLES_PER_SITE = 5;
export async function recordFailureSample(ctx: ServerContext, siteId: string, s: AdapterSample): Promise<void> {
await ctx.db
.insertInto("site_failure_sample")
.values({
site_id: siteId,
kind: s.kind,
request_kind: s.requestKind,
url: s.url,
http_status: s.httpStatus,
headers: JSON.stringify(s.headers),
message: s.message.slice(0, 500),
rule: s.rule,
body: s.body,
truncated: s.truncated,
created_at: new Date(s.capturedAt),
})
.execute();
await sql`delete from site_failure_sample where site_id = ${siteId} and id not in (
select id from site_failure_sample where site_id = ${siteId} order by created_at desc, id desc limit ${SAMPLES_PER_SITE})`.execute(
ctx.db,
);
}
/** 留样回调:不等写库,失败只记日志(不影响站点请求本身的结果) */
export function failureSampleSink(ctx: ServerContext, siteId: string): (s: AdapterSample) => void {
return (s) => {
void recordFailureSample(ctx, siteId, s).catch((err) =>
ctx.logger.warn({ siteId, err: { message: (err as Error).message } }, "站点失败页面留样写入失败"),
);
};
}
export async function listFailureSamples(ctx: ServerContext, siteId: string): Promise<SiteFailureSampleSummary[]> {
const rows = await ctx.db
.selectFrom("site_failure_sample")
.select([
"id",
"site_id",
"kind",
"request_kind",
"url",
"http_status",
"headers",
"message",
"rule",
"truncated",
"created_at",
sql<number>`octet_length(body)`.as("body_bytes"),
])
.where("site_id", "=", siteId)
.orderBy("created_at", "desc")
.orderBy("id", "desc")
.limit(SAMPLES_PER_SITE)
.execute();
return rows.map((r) => toApi<SiteFailureSampleSummary>(r));
}
export async function getFailureSample(
ctx: ServerContext,
siteId: string,
sampleId: string,
): Promise<SiteFailureSample | null> {
const row = await ctx.db
.selectFrom("site_failure_sample")
.selectAll()
.where("site_id", "=", siteId)
.where("id", "=", sampleId)
.executeTakeFirst();
return row ? toApi<SiteFailureSample>(row) : null;
}