[backport -> release/3.9.x] fix(plugin): fix ai latency metric doesn't correctly calculated issue

[team-gateway-bot]

Automated backport to release/3.9.x, triggered by a label in #14639.

Original description

Summary
Fix the prometheus plugin flaky test cases on AI metrics.

Root Cause:
AI metrics latency are floor evaluation, which means it could be 0, but AI metrics latency histogram only observe when latency is larger than 0, which may lead to missing increment of histogram.

Solution:

1. like other non-AI latency histogram, latency 0 should also be observed.
2. for test cases, Add waiting logic to wait the new latency calculation done.

Summary

Checklist

• The Pull Request has tests
• A changelog file has been created under changelog/unreleased/kong or skip-changelog label added on PR if changelog is unnecessary. README.md
• There is a user-facing docs PR against https://github.com/Kong/docs.konghq.com - PUT DOCS PR HERE

Issue reference

Fix KAG-7113