Rewrite offline-database article with modern sync section and improvement guide

[pubkey]

This PR contains:

• IMPROVED DOCS

Describe the problem you have without this PR

The articles/offline-database.md page was outdated. Its sync section described the replication only vaguely (and mentioned PouchDB), the content was thin on what an offline database is, and it gave no guidance on how to improve a setup beyond the basics. This PR fully rewrites the page:

• New structure: what an offline database is (vs a cache), why apps need one, RxDB intro, setup example, sync, encryption, improvement guide, FAQ, Follow Up.
• The sync section now describes the modern RxDB Sync Engine: the git-like document-level protocol, the three backend endpoints (pull handler, push handler, pull stream), checkpoint iteration and event observation modes, client-side conflict resolution, batched transfers, multi-tab leader election, and the current replication plugin list (HTTP, GraphQL, WebSocket, CouchDB, Firestore, MongoDB, Supabase, Appwrite, NATS, WebRTC/P2P, RxServer). The code sample uses realistic fetch-based pull/push handlers instead of empty placeholders.
• A new "How to Improve Your Offline Database Setup" section with eight concrete steps: storage choice, partial sync, cleanup of deleted documents, key compression, worker/shared-worker storages, schema migrations, deliberate conflict handling, and knowing the tradeoffs.
• The FAQ is rewritten to target real search queries with verdict-first answers.

The docs build (npm run docs:build) passes with no broken links on this page.

Todos

• Documentation

---

Generated by Claude Code