docs(tracing): add ChromaDB tracing pages

[SuhaniNagpal7]

Summary

• Adds TraceAI ChromaDB pages on both docs surfaces, following the 5-section sibling template (Install → Env vars → register → instrument → run).
• Introduces a new Vector Databases group in both sidebars and both overview grids. ChromaDB is the first entry; the group is structured so other Python vector-DB instrumentors (pinecone, qdrant, weaviate, milvus, lancedb, pgvector, redis-vector, mongodb-vector) can be added incrementally.
• Code example mirrors the runtime portion of the upstream README at traceAI/python/frameworks/chromadb/README.md verbatim (collection setup, add(...), query(...) with n_results=5), keeping a trailing print(results) for sibling-style consistency.

What changed

File	Change
src/pages/docs/tracing/auto/chromadb.mdx	New auto-instrumentation page
src/pages/docs/integrations/traceai/chromadb.mdx	New integrations page (mirrors content, distinct title/description)
src/lib/navigation.ts	+1 `Vector Databases` group with ChromaDB in each of the Tracing and Integrations sidebars
src/pages/docs/integrations/index.mdx	+1 `Vector Databases` card group
src/pages/docs/tracing/auto/index.mdx	+1 `Vector Databases` card group

Verification

• astro dev compiled both new MDX pages with zero errors / zero browser console errors. Routes `/docs/tracing/auto/chromadb` and `/docs/integrations/traceai/chromadb` both return `200`. Both overview grids and both sidebars resolve to the surface-appropriate URL.
• Live runtime smoke test against `api.futureagi.com`: ran the literal Step 2-5 code with `FI_API_KEY`/`FI_SECRET_KEY` from env. `register()` succeeded, `ChromaDBInstrumentor` instrumented, `collection.add(...)` + `collection.query(...)` ran, and 2 spans landed in the `CHROMADB_APP` project in the FAGI dashboard with the full documented `db.vector.*` attribute set (`db.system`, `db.namespace`, `db.operation.name`, `db.vector.query.top_k`, `db.vector.query.type`, `db.vector.query.include`, `db.vector.results.count`, `db.vector.results.ids`, `db.vector.results.scores`).

Heads-up (upstream-package bug, not a docs bug)

The `traceai-chromadb` instrumentor imports `wrap_function_wrapper` from `wrapt` and calls it with kwargs (`module=`, `name=`, `wrapper=`). That kwarg signature was removed in `wrapt` 2.0, so a fresh `pip install traceAI-chromadb` against today's PyPI wrapt (2.x) crashes with `TypeError: wrap_function_wrapper() got an unexpected keyword argument 'module'`. Pinning `wrapt<2` works around it. The right fix is in the upstream `traceAI` repo (either pin `wrapt<2` in `traceai-chromadb`'s `pyproject.toml` or switch the calls to positional form). Docs intentionally kept clean of the pin so they age well once the upstream is fixed.

Test plan

• `astro dev` and confirm `/docs/tracing/auto/chromadb` and `/docs/integrations/traceai/chromadb` both render with the 5 sections.
• Confirm a Vector Databases group with a ChromaDB entry appears in both sidebars (Tracing and Integrations), between Voice & Realtime and Java.
• Confirm the Vector Databases card group with a ChromaDB card appears on both overview grids.
• Optional: copy-paste the Step 2-5 code with real keys and confirm a trace lands in your FAGI workspace.