fix(openrouter): Avoid double-counting callModel metrics

[Ronald Koh (ronaldkohhh)]

Summary

• Stop logging aggregate token/cost metrics on OpenRouter callModel parent spans when child turn spans were emitted.
• Preserve parent output and metadata, including turn_count, while leaving per-turn usage on openrouter.beta.responses.send spans.
• Apply the same behavior to both OpenRouter and OpenRouter Agent instrumentation.

Why

OpenRouter callModel traces currently include usage metrics on both the aggregate parent span and the nested per-turn LLM spans. Span-level token and cost monitors then sum both levels and can report roughly double the actual usage for single-turn calls.

When no child turn spans are emitted, the parent still records aggregate metrics so usage is not dropped for simpler getResponse() flows.

Validation

• pnpm exec prettier --write js/src/instrumentation/plugins/openrouter-agent-plugin.ts js/src/instrumentation/plugins/openrouter-plugin.ts js/src/instrumentation/plugins/openrouter-agent-plugin.test.ts js/src/instrumentation/plugins/openrouter-plugin.test.ts
• pnpm exec vitest run src/instrumentation/plugins/openrouter-agent-plugin.test.ts src/instrumentation/plugins/openrouter-plugin.test.ts
• pnpm run check:typings

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6c954e3026

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "Codex (@codex) address that feedback".