1. Home
  2. Fetchie
  3. Indexing
  4. Troubleshooting

Troubleshooting

Most Fetchie problems come down to one thing: something standing between the indexer and your rendered pages. This page starts with the one-click diagnosis, then works through the usual suspects.

Start Here: the Test Page Fetch

On Tracksies > Fetchie > Index, click Test page fetch. Fetchie fetches your homepage exactly the way the indexer fetches every page — anonymously, as a logged-out visitor — and tells you what came back:

  • “Rendered pages fetch fine — N words read from your homepage” — the fetch path is healthy. If search results still look thin, the issue is elsewhere on this page
  • A warning that rendered pages can’t be fetched — the indexer is being blocked. Read on

Run this before rebuilding, not after. A rebuild against a blocked fetch path wastes the whole pass on thin content.

The Thin-Content List

When documents couldn’t be indexed from their rendered pages, the Index tab shows a warning listing them: they were indexed from stored post content instead, which is thin for builder-built pages — their body text may be missing from search. The list is your map of what needs re-fetching. Fix the cause below, then click Rebuild index to retry.

Why Rendered Fetches Fail

Three causes account for nearly all of it:

  • Maintenance or coming-soon mode. Logged-out visitors get a placeholder page, and the indexer is a logged-out visitor. Finish launching, or allowlist the indexer (next section), then rebuild
  • A security plugin or firewall rule blocking automated requests, or rate-limiting bursts of requests from the site’s own address
  • Blocked loopback requests. Some hosts stop a site from making requests to itself. Your host can confirm and usually fix this in minutes — it also affects WordPress’s own cron and site health checks, so they’ll know exactly what you mean

Individual pages can also fail for their own reasons — a page that redirects, errors, or responds with something other than a normal page won’t index from render. Those show up in the thin-content list too.

The search-bot tip

Fetchie’s indexer identifies itself politely: its requests carry the user agent Fetchie-Indexer. Most maintenance-mode and security plugins have an allowlist for search engine crawlers and named bots — add the Fetchie indexer there and it can read your pages even while real visitors see the coming-soon screen. That lets you build the index before launch day.

“My Insights Numbers Look Low”

Usually not a bug — three deliberate choices:

  • Administrator searches are never logged. Your own testing doesn’t count, on purpose
  • Page two isn’t a new search. Paging through results counts once
  • Empty-after-cleanup queries aren’t logged — a search of only filler words or single characters records nothing

If you’re comparing against an analytics tool that counts every results-page view, expect Fetchie’s number to be lower and more honest.

The Digest Didn’t Arrive

Quiet week (zero searches sends no email, by design), toggle off, or plain email delivery. The weekly digest page has the checklist.

The Debug Log

When you need to see what actually happened, Fetchie can narrate its work:

  1. Go to Tracksies > Settings > Debug
  2. Switch on logging for Fetchie
  3. Reproduce the problem (rebuild, wait for the daily pass, or save the affected page)
  4. View the log from the same screen

Fetch failures land there with the reason recorded — what URL failed and why (a timeout, an error code, a non-page response). Rebuild batches, the daily freshness pass, digest sends and skips are all summarised too. Logs are written per-plugin and cleaned up automatically when you switch logging off.

If a rebuild stops with “Indexing failed — check the browser console and try again”, the browser console note is literal: the error that stopped the batch is printed there, and it’s usually a security rule or an exhausted server timeout.

Related

How can we help?