# Scaling

# Scaling

Scaling Curiosity Workspace is about maintaining good user experience as your data size, query volume, and workload complexity increase.

# Scale starts with design

The biggest scaling wins are often design decisions:

• stable keys and idempotent ingestion (avoid duplication)
• correct schemas and edges (avoid expensive “workaround” queries)
• indexing only high-value fields
• using facets and graph constraints to reduce result sets

# Operational scaling areas

# Data volume growth

• plan storage growth (graph + indexes)
• implement incremental ingestion (avoid full refreshes)
• schedule reindex/reparse thoughtfully

# Query performance

• measure slow queries (graph and search)
• prefer bounded traversals and pagination
• avoid “fan-out” traversals without constraints

# Ingestion throughput

• batch writes and commit in chunks
• parallelize safely if supported by your connector architecture
• monitor failure rates and retries

# Practical scaling checklist

• Use hybrid search for better retrieval without excessive query logic.
• Prefer related facets and graph constraints to keep search scoped.
• Cache expensive aggregates (via endpoints) when used frequently.
• Separate environments and promote changes (schema, indexes) with review.

# Next steps

• Optimize relevance and retrieval: Search Optimization
• Build safe endpoint-based caching: Custom Endpoints