# Changelog of Kolesa.kz Scraper — Kazakhstan Car Listings & Prices (`logiover/kolesa-kz-scraper`) Actor

- **URL**: https://apify.com/logiover/kolesa-kz-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/logiover/kolesa-kz-scraper.md

## Changelog

### 2026-09-01

- Fixed runs being killed for running out of memory. A Kolesa listing page is ~900 KB of markup and the whole of it was being turned into a document just to read the twenty result cards, which took peak memory to the 512 MB ceiling. Cards are now parsed one at a time, cutting peak memory from ~490 MB to ~210 MB. Output is unchanged, field for field.
- Faster too: listing pages are now fetched a few at a time instead of strictly one after another, and the lighter parsing costs less CPU. The default 500-row run went from ~156 s to ~36 s.
- A page that cannot be fetched is now skipped rather than ending the crawl — one unlucky exit IP used to discard every remaining page.
- The time budget is now a real ceiling: a request is not started, and a retry not attempted, when there is not enough time left to finish it. Previously a retry chain begun just under the budget could run past the platform limit with rows still unsaved.
- A transient Apify Proxy error no longer aborts the run before the first page is fetched.
- The "page returned no cards" branch now says so in the log; it used to stop the crawl silently, which made a short run impossible to explain.

### 2026-08-11

- Maintenance release: refreshed the build and dependencies.
- Re-verified live execution, non-empty structured output and dataset field/type integrity.
- Reviewed reliability (retries, pagination) and output quality as part of a full-fleet QA pass.
