Introduce AGG across the docs

[em3s]

Background

AGG (real-time aggregation) — exposed via GET /edges/agg/{group} and supported across the engine — is partially documented. api-references/metadata.mdx describes Group and aggregation types (SUM, COUNT), and guides/build-your-commerce-app-with-live-fomo-counters.mdx exists as a placeholder labeled "Coming Soon". The concept, however, has not been introduced across the rest of the docs at the same level as GET / SCAN / COUNT. Readers landing on Design or top-level pages cannot tell that AGG exists or how it composes with the rest of the model.

The same gap holds on the storytelling side. AGG already powers production surfaces such as live "N people viewing" counters in KakaoTalk shopping, but those production cases are not yet reflected in stories/.

A subtle but important design fact: in the engine model, an Aggregator is not only a standalone query (/edges/agg/{group}) but also a modifier that attaches to Get, Count, Scan, and Seek queries. Documentation should reflect both shapes.

This umbrella tracks the documentation work needed to introduce AGG consistently — technical reference, user guides, and production storytelling.

Task

Design

• design/concepts.mdx: Introduce real-time aggregation as a core capability
• design/glossary.mdx: Add AGG, Aggregator, Group, and time bucket entries
• design/schema.mdx: Describe how aggregation groups are declared (fields, type, valueField, ttl)
• design/query.mdx: Add the standalone AGG query (pre-computed structures table, Query Types entry, Query Flow diagram branch)
• design/query.mdx: Document the Aggregator modifier attached to Get / Count / Scan / Seek
• design/mutation.mdx: Extend the Write-Time Optimization table and the closing line to include the AGG-side pre-computed structure
• design/storage-backends.mdx: Note aggregation-side storage implications, if relevant

API references

• api-references/query.mdx: Document the GET /edges/agg/{group} endpoint
• api-references/metadata.mdx: Cross-reference Group to the new query and design pages

Internals

• internals/encoding.mdx: Document AGG encoding (time buckets, group keys, value-field layout)

Top-level

• introduction.mdx: Surface AGG in the capability overview
• quick-start.mdx: Add a small AGG example after the GET / SCAN / COUNT walkthrough
• faq.mdx: Cover common AGG questions (windowing, group definitions, freshness)

Guides

• guides/build-your-commerce-app-with-live-fomo-counters.mdx: Complete the placeholder guide with concrete examples

Stories

• stories/use-cases/: Add a production case for live "N people viewing" counters in KakaoTalk shopping
• Cross-link existing stories that already touch aggregation (hbase-consistency.mdx, kakaotalk-gift-wish.mdx)

Done When

• AGG appears in Design, API references, and Internals at the same level of detail as GET / SCAN / COUNT
• A reader landing on introduction.mdx or quick-start.mdx can discover AGG without leaving the top-level pages
• The FOMO counters guide is no longer marked Coming Soon
• A production story for AGG is published under stories/use-cases/
• Sidebar order and section grouping are reviewed so the new pages slot in cleanly
• Korean translations under docs/ko/ are picked up by the translation pipeline (see Notes)

Notes

• This issue tracks documentation only. AGG itself is already implemented and in production.
• Confirm the translation pipeline auto-picks up new English pages; if manual, split that work into a follow-up issue.