chii
983259836d
docs: 联网搜索服务架构方案全套（plan-final/design-arch/选型决策/整合导览/MCP文档/部署预设/联调手册） bench: 5 方案 + 代理 + 站点矩阵本机实测工程（无密钥） 部署目标：primary mgr1 先行测试（待批准后执行）
124 lines
4.3 KiB
JSON
124 lines
4.3 KiB
JSON
{
|
||
"scheme": "trafilatura-http",
|
||
"group": "domestic",
|
||
"image": {
|
||
"ref": "bench-s3b-trafilatura:local",
|
||
"digest": "sha256:616c67688fe116adc223b45123b7eaf1275f3d5f9d8bfe4a746497b5855cd99f",
|
||
"base": "docker.io/library/python:3.12-slim-bookworm",
|
||
"base_digest": "python@sha256:f2431a8cca8c5c6b04bc1309ab7ce99cc36bda0e1787e88a7fac21d9b450a923",
|
||
"base_id": "sha256:f2431a8cca8c5c6b04bc1309ab7ce99cc36bda0e1787e88a7fac21d9b450a923",
|
||
"architecture": "arm64",
|
||
"base_architecture": "arm64",
|
||
"trafilatura": "2.2.0",
|
||
"inspect_line": "/bench-s3b-trafilatura sha256:616c67688fe116adc223b45123b7eaf1275f3d5f9d8bfe4a746497b5855cd99f bench-s3b-trafilatura:local"
|
||
},
|
||
"probed_at": "2026-09-01T11:44:02+08:00",
|
||
"rss_source": "cgroup",
|
||
"idle_mb_p50": 33.0,
|
||
"session_delta_mb": 2.18,
|
||
"burst_peak_mb": 40.61,
|
||
"session_peak_mb": 35.18,
|
||
"templates": {
|
||
"T2": {
|
||
"n": 5,
|
||
"success": 5,
|
||
"http_codes": [
|
||
200,
|
||
200,
|
||
200,
|
||
200,
|
||
200
|
||
],
|
||
"p50_s": 0.0585,
|
||
"p95_s": 0.1215,
|
||
"assert_ok": true,
|
||
"note": "T2×5 on T1 URL https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html (GET 200). Plan fallback www.gov.cn/zhengce/2021-03/13/content_5592681.htm returned HTTP 404 on 2026-09-01. Cross www.gov.cn/lianbo/202602/content_7057577.htm: host+container GET 200 (22354B) but extract(favor_precision=True)=None; recall/default only 62 chars of nav chrome.",
|
||
"url": "https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html"
|
||
},
|
||
"T5": {
|
||
"n": 60,
|
||
"success": 60,
|
||
"p50_s": 0.1152,
|
||
"p95_s": 0.4108,
|
||
"assert_ok": true,
|
||
"note": "60 tasks, client xargs -P 8 + server semaphore=8. 60/60 assert_ok, 0×429/503, OOM=false, wall 2s, all char_count=1122."
|
||
}
|
||
},
|
||
"t5": {
|
||
"n": 60,
|
||
"parallelism": 8,
|
||
"queue_impl": "server_semaphore_8+client_xargs_P8",
|
||
"success": 60,
|
||
"p50_s": 0.1152,
|
||
"p95_s": 0.4108,
|
||
"oom": false,
|
||
"http_429": 0,
|
||
"http_503": 0,
|
||
"first_ttfb_s": 0.0411,
|
||
"last_t_total_s": 0.4229
|
||
},
|
||
"t4_waf": [],
|
||
"scorecard": {
|
||
"mem": 100.0,
|
||
"quality": 100,
|
||
"stability": 100,
|
||
"latency": 85.0,
|
||
"ops": 90,
|
||
"total": 96.8,
|
||
"mem_idle": 100,
|
||
"mem_session": 100,
|
||
"mem_burst": 100,
|
||
"success_pct": 100.0
|
||
},
|
||
"t2_url": "https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html",
|
||
"pull": {
|
||
"image": "docker.io/library/python:3.12-slim-bookworm",
|
||
"source": "docker.io",
|
||
"mirror_used": false,
|
||
"started_at": "2026-09-01T03:31:14Z",
|
||
"finished_at": "2026-09-01T03:31:35Z",
|
||
"cmd": "docker pull --platform linux/arm64 docker.io/library/python:3.12-slim-bookworm",
|
||
"digest_from_pull": "sha256:f2431a8cca8c5c6b04bc1309ab7ce99cc36bda0e1787e88a7fac21d9b450a923",
|
||
"architecture": "arm64"
|
||
},
|
||
"commands": [
|
||
{
|
||
"ts": "2026-09-01T03:34:23Z",
|
||
"cmd": "docker pull --platform linux/arm64 docker.io/library/python:3.12-slim-bookworm"
|
||
},
|
||
{
|
||
"ts": "2026-09-01T03:34:25Z",
|
||
"cmd": "bash bench/trafilatura-http/run.sh up # compose build+up"
|
||
},
|
||
{
|
||
"ts": "2026-09-01T03:34:49Z",
|
||
"cmd": "bash bench/trafilatura-http/run.sh idle"
|
||
},
|
||
{
|
||
"ts": "2026-09-01T03:42:45Z",
|
||
"cmd": "bash bench/trafilatura-http/run.sh t2"
|
||
},
|
||
{
|
||
"ts": "2026-09-01T03:43:17Z",
|
||
"cmd": "bash bench/trafilatura-http/run.sh t5"
|
||
},
|
||
{
|
||
"ts": "2026-09-01T03:44:02Z",
|
||
"cmd": "T2_URL=https://www.gov.cn/lianbo/202602/content_7057577.htm bash request.sh t2 90"
|
||
}
|
||
],
|
||
"combo_rss_mb": {
|
||
"n": 10,
|
||
"p50": 131.74,
|
||
"min": 131.23,
|
||
"max": 132.0,
|
||
"cmd": "STATS_DIR=bench/searxng-cn/raw bash bench/searxng-cn/lib/sample_stats.sh p0-combo 1 30 bench-s3b-searxng-cn bench-s3b-trafilatura",
|
||
"sampled_at": "after dual idle, before T1/T2"
|
||
},
|
||
"notes": [
|
||
"no proxy; domestic direct fetch",
|
||
"plan fallback 十四五 URL HTTP 404; used T1 chinatax.gov.cn article",
|
||
"cross T2 www.gov.cn/lianbo/202602/content_7057577.htm: fetch 200 but favor_precision empty; recall yields nav-only 62 chars (首页|简|繁|EN). Metadata title readable: 三部门明确跨境电子商务出口退运商品税收优惠政策",
|
||
"image bench-s3b-trafilatura:local 312MB uncompressed; base python:3.12-slim-bookworm 214MB; trafilatura==2.2.0 from PyPI (no mirror needed)"
|
||
]
|
||
}
|