Add OpenTelemetry histogram metrics for LLM operations in aliyun_trace

[Copilot]

Summary

Implements 7 histogram metrics for LLM operations in the aliyun_trace component, exported via OTLP to Aliyun endpoint. All metrics include app_name tag.

Metrics Added:

• gen_ai.client.time_to_first_token - First token latency
• gen_ai.client.time_per_output_token - Per-token generation time (auto-calculated from duration/completion_tokens)
• gen_ai.client.time_between_token - Inter-token timing
• gen_ai.client.operation - Operation count
• gen_ai.usage.usage.prompt_tokens_details.cached_tokens - Cached token usage
• gen_ai.client.operation.duration - Total operation latency
• gen_ai.client.token.usage - Total token consumption

Implementation:

• MetricsClient (traceclient.py): OpenTelemetry SDK wrapper with histogram instruments, periodic export (5s), and graceful shutdown
• Integration (aliyun_trace.py): Metrics recorded automatically in build_workflow_llm_span() and message_trace() when LLM spans are created
• Tests (test_metrics_client.py): Unit tests covering initialization, recording, auto-calculation, and app_name tag validation

Metrics are recorded transparently when LLM operations execute in workflows or message-based apps. No configuration changes required beyond existing aliyun_trace setup.

# Metrics recorded automatically from usage data
usage_data = {
    "prompt_tokens": 100,
    "completion_tokens": 50, 
    "total_tokens": 150,
    "latency": 2.0
}
# → All 7 metrics exported with app_name tag

Screenshots

N/A - Backend metrics implementation

Checklist

• This change requires a documentation update, included: Dify Document
• I understand that this PR may be closed in case there was no previous discussion or issues. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.
• I ran dev/reformat(backend) and cd web && npx lint-staged(frontend) to appease the lint gods

Original prompt

在aliyun_trace组件中增加gen_ai.client.time_to_first_token、gen_ai.client.time_per_output_token、gen_ai.client.time_between_token、gen_ai.client.operation、gen_ai.usage.usage.prompt_tokens_details.cached_tokens、gen_ai.client.operation.duration、gen_ai.client.token.usage这些metric指标埋点，指标类型都采用histogram类型，这些指标埋点是LLM类型的埋点，应该在aliyun_trace中处理，所有指标中都要有app_name这个tag

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.