fix: incorrect l7 metrics

[yuanchaoa]

This PR is for:

• Agent

fix: incorrect l7 metrics

Steps to reproduce the bug

Changes to fix the bug

• 问题是HTTPv1的请求和响应都能正常解析，但是HTTP响应没有生成对应的L7Stats

Affected branches

• main
• 6.6

Checklist

• Added unit test to verify the fix.
• Verified eBPF program runs successfully on linux 4.14.x.
• Verified eBPF program runs successfully on linux 4.19.x.
• Verified eBPF program runs successfully on linux 5.2.x.
  =

[yuanchaoa]

@claude review一下这个提交

[claude]

Claude finished @yuanchaoa's task in 30m 6s —— View job

---

I'll analyze this and get back to you.

[yuanchaoa]

@claude review一下这个提交

[claude]

Claude Code is working…

I'll analyze this and get back to you.

View job run

[yuanchaoa]

@claude review一下这个提交，问题是HTTPv1的请求和响应都能正常解析，但是HTTP响应没有生成对应的L7Stats；看一下这个提交是否能解决这个问题

[claude]

Claude Code is working…

I'll analyze this and get back to you.

View job run

[kylewanginchina]

@claude review一下当前代码

[claude]

Claude finished @kylewanginchina's task in 3m 9s —— View job

---

Todo

• Read CLAUDE.md and repo guidelines (none found)
• Review PR diff and relevant code
• Provide review feedback

Summary

Reviewed the change in agent/src/common/l7_protocol_log.rs:352-362 that relaxes request/response ordering from strict </> to <=/>=.

Review

• ✅ This looks correct for the reported bug: when request/response share the same timestamp (coarse time resolution or batching), the prior strict comparison would skip pairing and drop response stats. Allowing equality lets the pair be matched and still computes rrt = 0 as expected.
• 💡 Optional: consider adding a small regression test that constructs two LogCache entries with identical time and verifies L7PerfStats counts include the response. This would lock in the fix.