Handling 500k+ SKU catalogs in Magento, attribute architecture, search performance

Yes, Magento handles 500k+ SKUs in production. The biggest industrial catalog I’ve shipped runs ~620k SKUs and faceted-search filters return under 200ms on mobile.

The architecture choices that matter:

• EAV attribute sets per category family, fasteners get thread, drive type, material, grade, ANSI/DIN spec. Bearings get bore, OD, width, dynamic load, ABMA class. Motors get HP, RPM, frame, voltage, NEMA enclosure. Don’t share one attribute set across all 500k SKUs; that’s how you end up with a 600-column flat table that nothing queries fast.
• OpenSearch 2.x replaces MySQL search + Elasticsearch 7. Magento 2.4.6+ supports OpenSearch natively. With compound aggregations and per-attribute keyword analyzers, sub-200ms faceted search across 500k SKUs is realistic.
• Stock-status denormalised per warehouse. The native price/stock indexers slow at 100k+ SKUs across multiple warehouses; a custom denormalised table for “in-stock at warehouse X” filtering keeps category-page filtering fast.
• Catalog flat tables off, EAV indexed cleanly. Counter to what you’d expect, flat tables hurt at 500k+ SKUs because the row width destroys MySQL page cache hit rate.
• Image CDN-served, lazy-loaded. Industrial catalogs often have 3-8 images per SKU plus a CAD download. Cloudflare / Fastly / Cloudinary for images; CAD files on S3 with signed URLs.

Beyond architecture, the operational piece is catalog import pipelines. Industrial SKUs come from supplier PDFs, EDI 832 catalogs, distributor PIM exports. Akeneo PIM or a custom Magento import module is mandatory; nobody hand-enters 500k SKUs.