Use Case · Commerce

Catalog search on the Software Lifecycle Cloud

Catalog search on KERNSTACK is a text index declared over the products table itself: titles, descriptions, categories, and the attributes shoppers filter by, with synonyms maintained as merchandiser-owned data. Search results read live catalogue rows, so price and availability are never the stale copies a separate search cluster serves.

The engineering bottleneck with assembled stacks

Merchandisers fight a search that lives somewhere else: the index lags the catalogue by a sync cycle, the synonym list that would fix “sofa” versus “couch” is a ticket to whoever owns the search cluster, and the Black Friday postmortem discovers the bestseller was unfindable because of a field that stopped syncing in October.

Architecture on the Software Lifecycle Cloud

The products table is the only catalogue; the text index is declared over its searchable fields and rebuilds from it. The search function combines index hits with live reads — facet counts, stock states, prices — so merchandising rules (boost this season’s line, bury out-of-stock) are code and data in the project, reviewable before the sale starts, not a vendor console’s drag-and-drop.

Synonyms compound like capital

Every “customers call it X” learned by support becomes a synonym row, and the index gets permanently smarter in the direction of your customers’ vocabulary. That asset normally accumulates inside a search vendor’s configuration, unexportable; here it is a table beside the products, as portable and reviewable as the catalogue itself.

  • Index derives from the live table — new products are findable when they exist, not after the nightly sync
  • Facets and boosts evaluate against live rows, so price and stock in results are the truth
  • Zero-result queries log to a table: the merchandising agenda, written by customers

When a search specialist earns its invoice

Learning-to-rank over massive behavioural data and multi-million-SKU catalogues are specialised work. The honest majority — thousands of products, filters, synonyms, seasonal boosts — is an index declaration and merchandising discipline, which this stack keeps inside the business that owns the catalogue.

Frequently asked questions

Candidate retrieval comes from the declared text index; facets, ranges, and stock filtering apply against live product rows in the search function. That split keeps results honest — the index proposes, the record disposes.

Merchandisers, as data. The synonym table is edited through your admin surface and the change carries an author and a date. Zero-result query logs tell them exactly which synonym to add next.

The index is declared over the products table and updates from it, so a price change or a new product does not wait for a sync window. Launch-morning catalogue edits are findable on launch morning.

Related capabilities, architectures, and comparisons

Ship your stack on the Software Lifecycle Cloud.

Closed preview on dedicated European infrastructure. Bring a repository or start from a blueprint.