Slow-Page Safety Retry

Slow-Page Safety Retry is a single longer rendering attempt RankGear makes after a page returns an empty hard timeout or a provisional load failure, run within the same bounded safety limits so a genuinely slow page still gets one fair chance to render.

TermSlow-Page Safety Retry
CategoryAPIs, Reliability and Publishing
Also known asLonger Page Retry
Where it appearsAnalysis Run, On-Page Audit, and Bulk Runs

What it means in RankGear

When RankGear fetches and renders a page, it works under a fetch budget so a single unresponsive URL cannot stall a whole run. If a page hits that hard ceiling with nothing usable returned, or the initial load fails only provisionally, the run does not immediately mark it dead. Instead it grants one Slow-Page Safety Retry: the same page is rendered again with a longer allowance, still capped by the established safety limits. You see the outcome reflected in the run status and per-URL diagnostics for Analysis Run, On-Page Audit, and Bulk Runs, where a page that succeeded on the second attempt is treated as a normal result rather than a failure.

How to interpret it

Read this as RankGear separating “slow” from “broken.” A retry firing tells you the first attempt timed out with no content, not that the page is misconfigured; a page that renders on the retry is fully valid data. What the retry deliberately does not do is loop. It is one longer attempt, not open-ended re-polling, so a URL that still returns nothing after the retry is a real fetch failure worth investigating rather than a timing fluke. When you are triaging a failed URL, check the visible status and diagnostics first: a transient slow render behaves differently from an authorization or credential problem, and only the former is what this retry is meant to rescue.

SituationWhat the retry does
Empty hard timeout or provisional load failureOne longer rendering attempt, within bounded limits
Page renders on the second attemptCounted as a normal, valid result
Still empty after the retryRecorded as a genuine fetch failure, no further loops
Authorization or credential failureStops without repeated paid calls

Example

You queue a Bulk Run across 40 competitor URLs. One of them is a heavy, script-laden landing page that does not finish rendering before the fetch budget expires, so the first attempt returns empty. RankGear applies a Slow-Page Safety Retry, gives that single URL a longer allowance, and the page renders in full the second time, so it lands in your report as ordinary data. In the same run a different URL sits behind an authorization wall; that one fails on the first response and is not retried, because repeating the request would only burn time and paid calls without changing the outcome.

Important considerations

  • The retry is bounded and singular by design: one longer attempt, not continuous re-polling, so it cannot let a slow site hang an entire run.
  • It targets timing and provisional load failures only. Authorization and credential errors are not retried, because the problem is configuration, not speed.
  • Where the retry touches credentials or remote publishing targets, tests should stay read-only where possible and secrets must remain redacted in logs and diagnostics.
  • A page that renders on the retry is data of the same quality as a first-attempt render; a page that fails after it is a real failure to investigate, not noise.

Related terms

Part of the RankGear glossary · how RankGear measures · the 870 factors.