From 3c3d382c8dd65f5b3c9fed2d907a03208ec7bd16 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Gerg=C5=91=20M=C3=B3ricz?= Date: Mon, 15 Sep 2025 20:59:01 +0200 Subject: [PATCH] feat(scrapeURL/index): index metrics (#2160) Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com> --- .../src/scraper/scrapeURL/engines/index/index.ts | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/apps/api/src/scraper/scrapeURL/engines/index/index.ts b/apps/api/src/scraper/scrapeURL/engines/index/index.ts index 538c33a1a..95eb1d0ae 100644 --- a/apps/api/src/scraper/scrapeURL/engines/index/index.ts +++ b/apps/api/src/scraper/scrapeURL/engines/index/index.ts @@ -12,11 +12,9 @@ import { addIndexInsertJob, generateDomainSplits, addOMCEJob, - addDomainFrequencyJob, } from "../../../../services"; import { EngineError, IndexMissError } from "../../error"; import { shouldParsePDF } from "../../../../controllers/v2/types"; -import { storage } from "../../../../lib/gcs-jobs"; export async function sendDocumentToIndex(meta: Meta, document: Document) { const shouldCache = @@ -300,6 +298,12 @@ export async function scrapeURLWithIndex( } if (selectedRow === null || selectedRow === undefined) { + meta.logger.debug("Index metrics", { + module: "index/metrics", + hit: false, + maxAge, + dynamicMaxAge: meta.options.maxAge === undefined, + }); throw new IndexMissError(); } @@ -333,6 +337,14 @@ export async function scrapeURLWithIndex( } } + meta.logger.debug("Index metrics", { + module: "index/metrics", + hit: true, + age: Date.now() - new Date(selectedRow.created_at).getTime(), + maxAge, + dynamicMaxAge: meta.options.maxAge === undefined, + }); + return { url: doc.url, html: doc.html,