In short. Open Crawl Stats, select a response group and inspect an example URL to see details of the individual crawl request.
Definition: Google crawl failures are unsuccessful attempts by Google to fetch URLs from a website. Search Console's Crawl Stats report groups crawl requests by response type and provides example URLs with details of individual requests.
Key facts
- As of October 2026, Google's Crawl Stats documentation describes response percentages based on request counts, including repeated requests for the same URL.
- Google's HTTP status documentation states that server errors in the 5xx range and 429 responses temporarily slow its crawlers.
- Google's URL Inspection documentation explains that a live test fetches and examines a URL in real time.
Where is the response breakdown?
Google's Crawl Stats report documentation gives this navigation path:
Property settings > Crawl stats
The report is available for Domain properties and root-level URL-prefix properties. Google's examples include https://example.com and http://m.example.com.
In the crawl response table, selecting a response group opens its detailed view and example URLs. Selecting an example opens details for that specific crawl request. The documentation lists crawl time, response code and response size among the available request details. For server errors, the thumbnail chart shows approximately when failures occurred; opening the details provides exact times.
Which response groups indicate failures?
Google's response-group descriptions identify these failure types:
| Response group | Meaning |
|---|---|
| Server error (5XX) | Server errors that cause availability warnings. |
| DNS unresponsive | The hostname lookup service fails to respond to requests. |
| Fetch error | A bad port number, IP address or unparseable response prevents fetching. |
| Page timeout | The page request times out. |
| Redirect error | Redirection fails, for example through a circular redirect or too many redirects. |
| Page could not be reached | The request fails before reaching the server, so it has no corresponding server-log entry. |
The same documentation treats Not found (404) as potentially appropriate: a deleted page without a replacement can correctly return that response. Permanent and temporary redirects appear among the report's good response codes.
When Host status is red, its details show robots.txt availability, DNS resolution and server connectivity. The robots.txt file contains crawling rules; its availability check concerns whether Google can retrieve that file.
How should the sample URLs and percentages be read?
Google's data and example-URL guidance says the listed addresses are the actual URLs requested by Google.
The response table measures requests. Google's example uses a URL requested twice: a Server error (500) followed by OK (200) produces 50% Server error and 50% OK. Each request in a server-side redirect chain also counts separately.
The example list is a representative sample that can be weighted by day; a URL absent from the list can still have been requested. For URL-prefix properties, totals include both HTTP and HTTPS requests, while example URLs use the property's specified protocol.
Related terms
- HTTP status codes
- Server log analysis
- How do I check host availability in Search Console Crawl Stats?
- How does Search Console's live URL test differ from indexed data?
FAQ
How can I check a failed example after a repair?
You can inspect the address in URL Inspection and select Test live URL. Google's URL Inspection guide says this fetches the page in real time, so the result can differ from previously indexed data.
Why can crawl totals exceed the requests in my server logs?
Google's Crawl Stats troubleshooting guidance explains that unavailable robots.txt files can cause potential crawls to count even though the requests never happen. You can check robots.txt fetching status to investigate this cause.
Sources
- Google Search Console Help, "Crawl Stats report", undated
- Google Search Central, "How HTTP status codes affect Google's crawlers", 4 February 2026
- Google Search Console Help, "URL Inspection tool", undated
Check how a page reads to answer engines. Open the GEO score analyzer