Noyaru
Log in Get started →

Troubleshooting: common problems

Empty or incomplete crawl, Search Console returning nothing, missing fix buttons, an issue that persists: the causes and what to check.

The cases that come up most often, with the first thing to check.

The crawl only finds the home page

Start by checking the links and the scope actually analyzed.

  • Are the links real <a href="…"> tags? A menu built in JavaScript without real links is not followed by anyone.
  • Is robots.txt blocking crawling? Open yoursite.com/robots.txt.
  • Do the page limit, the subdomains and the include/exclude filters cover the expected addresses? These choices are in the analysis settings.
  • Is the start URL really the canonical version — right protocol, right www? Otherwise the first hop is a redirect.

The crawl stopped before the end

Three things to check in the analysis tracking, available from Activity:

  1. Page limit reached — the project's, the plan's, or the monthly quota.
  2. Maximum analysis duration exceeded — a slow site, or a lot of pages.
  3. The host refused the requests — 403 or 429. Reduce the simultaneous connections in the crawl options, then review the tracking before running it again.

In the third case, these refusals are not counted as errors on your site: they appear as a crawl incident.

→ Run a crawl

Search Console is connected but no data comes in

First check that the property matches the project URL.

Check character by character: https versus http, with www versus without, trailing slash. If you have a Domain property (sc-domain:…), use it: it covers every case.

Other things to check:

  • Does the authorized Google account actually have access to this property?
  • Is the requested period wide enough? On a new site, 7 days may contain nothing.
  • Is the minimum impressions threshold excluding all the traffic of a small site?

Reminder: Google publishes its data with a two- to three-day delay.

→ Connect Search Console

The fix buttons do not appear

In the order you should check:

  1. Is a repository linked to this project? The banner on the Fixes page tells you.
  2. Is GitHub still connected? A token revoked on the GitHub side cuts the connection without warning.
  3. Can the issue be fixed in the code? A 5xx error or an editorial decision has no button, and that is intentional.
  4. Does your plan include AI fixes? The Free plan does not include them.

→ Connect GitHub

“Fix created” but nothing in my repository

Look at the pull request link on the task, on the Fixes page. It is almost always there, on a dedicated branch — and not on your main branch, which is the expected behavior.

If the PR exists but does not appear in your usual list, check that you are looking at the right repository and the right target branch.

The issue is still there after the merge

Three causes, in this order:

  1. The deployment did not happen, or it failed.
  2. A cache is still serving the old version — CDN, page cache, service worker.
  3. The verification crawl predates the deployment.

Quick test: open the URL in a private window and look at the source code. If the fix is there, it is a cache. If it is not, it is a deployment.

→ Verify a fix

A score dropped even though I changed nothing

Almost always explainable:

  • New pages were published and carry template issues.
  • The previous crawl was partial — canceled, or stopped at a limit. Comparing a full crawl to a partial crawl makes no sense.
  • A linked external resource went down, which makes broken links appear.

Never compare two scores: compare the two crawls. The comparison view names exactly what appeared.

→ Read the report

An issue I think is wrong

Open the issue's detail page and look at the evidence column: it shows exactly what the crawler received.

Two common cases where the disagreement is only apparent:

  • You are looking at the rendered page, the crawler read the served HTML. If a tag is injected by JavaScript afterward, it may be missing from the initial response.
  • You are logged in, the crawler is not. A page that changes depending on the session is not the same on both sides.

If the evidence still does not match what you are seeing, write to contact@noyaru.com with the URL and the issue in question: that is exactly the kind of case that improves the crawler.

Contact support

contact@noyaru.com, with four items: the account email, the project name, the approximate time of the problem, and a screenshot. They are almost always enough to trace the problem on the server side.

Tracking no longer updates

An interruption message keeps the last state received; it does not prove that the analysis has stopped. Click Refresh status before making any new request. If your session has expired, use the sign-in link: it keeps the screen you started from.

A job that finished without a readable report stays flagged as such. An older report available from the site must not be read as the result of the new request.

Frequently asked questions

Who should I write to if none of this solves my problem?

To contact@noyaru.com, including the account email, the project name, the approximate time of the problem and a screenshot. These four items are almost always enough to trace the problem on the server side.

Where can I see whether the service itself has an incident?

The Status page shows the state of the services. If a component is in error there, there is no need to look any further on your end.

Try Noyaru

Create a project, run a crawl and browse your first report. No credit card required.

Create a free account