kernel/bpf-developer-tutorial
1
0
Fork 0
mirror of https://github.com/eunomia-bpf/bpf-developer-tutorial.git synced 2026-04-13 17:09:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

Deploying to gh-pages from @ eunomia-bpf/bpf-developer-tutorial@633aac43f1 🚀

Browse Source
This commit is contained in:
yunwei37
2023-05-06 18:10:51 +00:00
parent bcc37b85d4
commit f56e271d19
12 changed files with 864 additions and 187 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
13-tcpconnlat/tcpconnlat.html
View File

@@ -264,28 +264,8 @@ cleanup:
return 0;
}
</code></pre>
<h3 id="eunomia-测试-demo"><a class="header" href="#eunomia-测试-demo">Eunomia 测试 demo</a></h3>
<p>使用命令行进行追踪：</p>
<pre><code class="language-bash">$ sudo build/bin/Release/eunomia run tcpconnlat
[sudo] password for yunwei:
[2022-08-07 02:13:39.601] [info] eunomia run in cmd...
[2022-08-07 02:13:40.534] [info] press 'Ctrl C' key to exit...
PID COMM IP SRC DEST PORT LAT(ms) CONATINER/OS
3477 openresty 4 172.19.0.7 172.19.0.5 2379 0.05 docker-apisix_apisix_1
3483 openresty 4 172.19.0.7 172.19.0.5 2379 0.08 docker-apisix_apisix_1
3477 openresty 4 172.19.0.7 172.19.0.5 2379 0.04 docker-apisix_apisix_1
3478 openresty 4 172.19.0.7 172.19.0.5 2379 0.05 docker-apisix_apisix_1
3478 openresty 4 172.19.0.7 172.19.0.5 2379 0.03 docker-apisix_apisix_1
3478 openresty 4 172.19.0.7 172.19.0.5 2379 0.03 docker-apisix_apisix_1
</code></pre>
<p>还可以使用 eunomia 作为 prometheus exporter，在运行上述命令之后，打开 prometheus 自带的可视化面板：</p>
<p>使用下述查询命令即可看到延时的统计图表：</p>
<pre><code class="language-plain"> rate(eunomia_observed_tcpconnlat_v4_histogram_sum[5m])
/
rate(eunomia_observed_tcpconnlat_v4_histogram_count[5m])
</code></pre>
<p>结果：</p>
<p><img src="tcpconnlat_p.png" alt="result" /></p>
<h3 id="编译运行"><a class="header" href="#编译运行">编译运行</a></h3>
<p>TODO</p>
<h3 id="总结"><a class="header" href="#总结">总结</a></h3>
<p>通过上面的实验，我们可以看到，tcpconnlat 工具的实现原理是基于内核的TCP连接的跟踪，并且可以跟踪到 tcp 连接的延迟时间；除了命令行使用方式之外，还可以将其和容器、k8s 等元信息综合起来，通过 <code>prometheus</code> 和 <code>grafana</code> 等工具进行网络性能分析。</p>
<blockquote>