Go · Search Implementation

Go with KERNSTACK search

Go handlers plus a declared text index is how product search ships on KERNSTACK without a search cluster: the index covers the fields your Go code searches, the function queries it and then reads the winning records live, and filters — stock, territory, permissions — apply against the rows themselves rather than a stale copy.

Full-Text Search stack.json IaC

Talk to us Read Technical Documentation → Software Lifecycle Cloud →

Go Search

The declaration does the architectural work: one index entry names its sources and fields, and the platform keeps it current with the table. Go code never runs an indexer, a syncer, or a reindex cron — the category of background job that quietly stops in month three of an assembled stack.

Query flow is deliberately two-step. The index proposes candidate ids by relevance; the handler fetches those records and applies the rules only live data can answer (is it in stock, may this user see it, what is today’s price). Search quality work — synonyms, boosts, zero-result review — happens against the logged queries in your own tables.

Costs stay legible: index storage at the storage rate, queries as function GiB-seconds. A catalogue doing two million searches a month spends a few hundred credits on query compute, which reframes most “we need a search platform” conversations.

CLI

Plan Go Search, then deploy.

Terminal — CLI workflow kernstack cli

kernstack add function --name api --blueprint func-go-minimal
kernstack plan
kernstack deploy

Docs for Go Search. Search on the platform.

go-search.txt Go

kernstack add function --name search
# declare search_indexes over the products table (below)
kernstack plan && kernstack deploy
kernstack logs --function search   # watch real queries while tuning

stack.json — Go + search plan · deploy

{
  "resources": {
    "functions": [{ "name": "search" }],
    "tables": [{ "name": "products", "key_name": "product_id", "key_type": "S", "indexes": [] }],
    "search_indexes": [{ "name": "catalog", "index_type": "text",
      "sources": [{ "type": "tables", "name": "products", "fields": ["title", "description", "tags"] }] }]
  }
}

FAQ

Frequently asked questions

Do we write indexing code in Go?

No. Indexing is the platform’s side of the declaration. Your Go code writes records and queries the index; nothing in the repository resembles an indexer.

Can search span documents as well as records?

Yes. Add a bucket source (optionally prefix-scoped) beside the table source in the same index, and one query covers structured records and document content.

How do we improve ranking over time?

From your own query logs: log searches and zero-result queries to a table, then adjust fields, synonyms (as data), and handler-side boosts. The feedback loop is project data, not a vendor dashboard.

Connected in the Project Graph

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.

Talk to us See pricing & credits →