⚡ Why materialized views? Real numbers

On a ~50k-row dataset, reading from pre-aggregated materialized views turns heavy joins into double-digit to triple-digit speedups compared to running the raw SQL each time.

All features are designed to be production-ready with following principles

• High availability: MVs are created and refreshed in the background, ensuring minimal downtime.
• Repeatable operations: Clear APIs and CLI tasks for consistent behavior.
• Observability: Track runs, errors, and performance metrics.
• Rails-native: Integrates seamlessly with Active Job, Rails logger, and error handling.
• Extensible: Supports multiple job adapters (ActiveJob, Sidekiq, Resque) and can be customized for specific needs.
• Security: Contributions to security are encouraged, with a dedicated policy for reporting vulnerabilities.
• Community-driven: Contributions are welcome, with a CLA to ensure legal clarity.
• All features are free and open source under the MIT license. There is no other version or paid tier.

Sample run (5 iterations)

With 50,000 rows

view	iterations	baseline(ms) min|avg|max	mv(ms) min|avg|max	speedup_avg
mv_user_accounts	5	16 | 31 | 74	1 | 2 | 5	15.5
mv_user_accounts_events	5	70 | 78 | 108	1 | 1 | 2	78.0
mv_user_activity	5	159 | 161 | 165	1 | 1 | 2	161.0
mv_user	5	1 | 1 | 2	1 | 2 | 7	0.5

Stability check (100 iterations)

With 50,000 rows

view	iterations	baseline(ms) min|avg|max	mv(ms) min|avg|max	speedup_avg
mv_user_accounts	100	15 | 17 | 69	1 | 1 | 20	17.0
mv_user_accounts_events	100	70 | 70 | 73	1 | 1 | 3	70.0
mv_user_activity	100	158 | 161 | 242	1 | 1 | 2	161.0
mv_user	100	1 | 1 | 1	1 | 1 | 2	0.5

Takeaways

• Multi-table aggregates shine: ~70× (accounts+events), ~161× (full activity).
• Single-table scans: little/no benefit; use normal indexes or caching.
• Materialize expensive joins/aggregations you read often.
• PostgreSQL
  • Materialized views (MVs) make it faster for complex queries, especially those involving expensive joins or aggregations.
  • MVs are not a silver bullet for all queries; use them when they fit the use case.
  • If you have a slow query with poor performance, MVs might help you speed it up significantly.
  • MVs are not a replacement for proper indexing and query optimization.
  • Read more about PostgreSQL materialized views.