Evaluate instrumentation for Files API

The OpenAI integration does not instrument file management endpoints. These are supporting operations (uploading training data, batch inputs, etc.) rather than inference calls.

**Endpoints to evaluate:**

* `openai.resources.Files.create()` (upload)
* `openai.resources.Files.retrieve()`
* `openai.resources.Files.list()`
* `openai.resources.Files.delete()`
* `openai.resources.Files.content()` (download)

**Decision needed:**
Determine whether lightweight instrumentation (file ID, purpose, size) would provide useful telemetry, or if this is out of scope for GenAI monitoring. Consider that file operations are ancillary to the core inference/generation use cases.

**Reference:** [OpenAI integration compliance review](<../openai-integration-compliance-review.md>), Section 6

從 ../openai-integration-compliance-review.md 的第 6 節開始，檢查圍繞 Files.create()、retrieve()、list()、delete() 和 content() 的 OpenAI 整合。評估檔案 ID、用途和大小遙測資料是否對這些操作有用。針對 Files API 端點，記錄明確的納入範圍或排除範圍決定及其理由，即視為完成。

由索引模型根據 Issue 內容生成。