Hacktoberfest 2026: los issues que los mantenedores marcaron para octubre, abiertos y aptos para principiantes. Explorar issues de Hacktoberfest

Marketing site search (Algolia) returns stale, unversioned docs links

Abierto
  • ¿Algo no está bien?
#617 0 comentarios 0 reacciones 0 asignados Ver en GitHub

Los mantenedores suelen responder en 1 día

Nadie ha tomado este issue todavía.

Evaluación

Dificultad
4/5
Tiempo estimado
3-5 días
Aptitud para principiantes
38/100
Tipo de issue
Error
Claridad
Bien especificado
Estado de actividad
Activo
Stack tecnológico
next.js, typescript
Área
search

Línea de trabajo

The issue points to the marketing search integration in src/components/navbar.tsx and the https://kagent.dev/docs/sitemap.xml sitemap. Start by checking the Algolia DocSearch crawler status and its start URLs or sitemaps; then unblock it, run a crawl, and accept the expected safety check. Done means searching “BYO OpenAI” returns a link to /docs/kagent/0.x/supported-providers/byo-openai/.

Escrito por el modelo de indexación a partir del texto del issue.

Descripción

bug DocOps

Problem

The search box on the kagent.dev marketing pages (Algolia DocSearch, in src/components/navbar.tsx) returns stale results:

  • Every result uses a pre-versioning URL, such as /docs/kagent/supported-providers/byo-openai. None include 0.x/ or 1.x/.
  • No 1.x pages appear at all.
  • Result text reflects the docs as of late July 2026.

The search inside /docs/ uses the Hugo FlexSearch index and is not affected.

Old URLs redirect to 0.x through next.config.mjs. Today, any old URL with two or more path segments returns a 500 instead of redirecting, which affects 52 of the 60 kagent pages in the index. That redirect bug is fixed separately. This issue covers only the stale index.

What the index shows

The public search key can list indexes, so the dates below come from Algolia's GET /1/indexes endpoint (app 0Q0AZY5UR3):

Index Last written Records
kagent (live, read by the site) 2026-07-31 1,934
kagent.tmp (crawler scratch index) 2026-08-21 1,479

The DocSearch crawler writes each crawl to kagent.tmp and swaps it into kagent only if it passes the crawler's checks.

Likely cause

This is a best guess, since the crawler dashboard isn't visible with the public key:

  1. On 2026-08-19, the old Next.js docs pages were removed (ad3a69d).
  2. The 2026-08-21 crawl found about 24% fewer records. The crawler's safety check blocks publishing when too many records disappear, so the live index was kept.
  3. A blocked crawler stays stopped until someone unblocks it. kagent.tmp hasn't been written since 2026-08-21, which fits.

The 2026-08-21 crawl ran against the Hugo pages and still got 1,479 records, so the existing selectors probably mostly work on the Hextra markup.

Fix

Someone with access to the Algolia crawler (DocSearch was added in e9e4b03):

  • Check the crawler status. Confirm it's blocked, or find the actual error.
  • Make sure the crawler's start URLs or sitemaps include https://kagent.dev/docs/sitemap.xml, which lists the 0.x and 1.x pages.
  • Unblock the crawler and run a crawl. Accept the safety check for this run, because a large record change is expected.
  • Spot-check results: search "BYO OpenAI" and confirm the result links to /docs/kagent/0.x/supported-providers/byo-openai/.

Note: the crawler probably skips pages marked noindex by default. While 0.x is the current release, 1.x pages that also exist under 0.x are marked noindex, so a healthy crawl may show few 1.x results. That's likely fine while 1.x is in alpha.

Longer term (optional)

The site has two search indexes for the same docs: Algolia on the marketing pages and FlexSearch under /docs/. Only the FlexSearch one rebuilds on every deploy. Consider either scheduling a crawler health check, or pointing the marketing search at the Hugo index and retiring Algolia.

Lenguaje dominante
TypeScript
Estrellas
21
Forks
69
Merge medio
14 h 16 min
PR fusionados (30 d)
105

Preparar el entorno

Este proyecto no incluye contenedor de desarrollo, Dockerfile ni guía de contribución, así que la configuración corre por tu cuenta: empieza por su README y consulta nuestra guía para la primera contribución para los pasos generales.

Primeros pasos

  1. Lee el issue completo y luego la guía de contribución del proyecto.
  2. Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
  3. Haz un fork del repositorio y trabaja en una rama.
  4. Abre un pull request que haga referencia al número del issue.

Más de kagent-dev/website

Todos los issues de kagent-dev/website

Issues similares

Más issues de TypeScript

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.