feat(attributes): Add legacy gen_ai cache token attributes (deprecated)

[JPeer264]

Adds gen_ai.usage.cache_creation_input_tokens and gen_ai.usage.cache_read_input_tokens, which the JavaScript SDK stopped emitting in v11 in favour of the dotted gen_ai.usage.cache_*.input_tokens names.

Both join the existing alias groups that already hold the Vercel AI spellings gen_ai.usage.input_tokens.cache_write and gen_ai.usage.input_tokens.cached.

Deprecated by getsentry/sentry-javascript#22811

[github-actions]

Semver Impact of This PR

🟡 Minor (new features)

📋 Changelog Preview

This is how your changes will appear in the changelog.
Entries from this PR are highlighted with a left border (blockquote style).

---

New Features ✨

Attributes

• Add legacy gen_ai cache token attributes (deprecated) by JPeer264 in #582

• Add graphql.processing.type by andreiborza in #572

Other

• (span-names) Add url.domainrule for http.client spans by Lms24 in #575

Bug Fixes 🐛

Names

• Add missing gen_ai ops to span name rules by Lms24 in #579
• Include tool name in gen_ai.execute_tool span names by Lms24 in #580
• Correct graphql span name examples by andreiborza in #573

Internal Changes 🔧

• (attributes) Add search alias for segment name by nsdeschenes in #577

---

_{🤖 This preview updates automatically when you update the PR.}

[vgrozdanic]

we need to first update the relay code for cost calculation before we can merge this, we currently never read gen_ai.usage.cache_creation.input_tokens in relay, and this change would make relay rewrite all the older fields to this new field gen_ai.usage.cache_creation.input_tokens.

This is a good change, we just need to first catch up with it in other places (and product potentially).

I see that OTEL has changed this last week: https://github.com/open-telemetry/semantic-conventions-genai/blob/67dff024110be5bd9f318006e733f4078e0f4c97/changelog.d/440.breaking.md?plain=1#L1

[vgrozdanic]

we need to first update the relay code for cost calculation before we can merge this, we currently never read gen_ai.usage.cache_creation.input_tokens in relay, and this change would make relay rewrite all the older fields to this new field gen_ai.usage.cache_creation.input_tokens.

This is a good change, we just need to first catch up with it in other places (and product potentially).

I see that OTEL has changed this last week: https://github.com/open-telemetry/semantic-conventions-genai/blob/67dff024110be5bd9f318006e733f4078e0f4c97/changelog.d/440.breaking.md?plain=1#L1

Relay will handle this quite well already, i wasn't aware that we don't have to worry about this, so this change is safe