Site Audit Troubleshooting

Use this page to diagnose and resolve problems with your Site Audit projects.

Common run issues

Problem	Possible cause	What to do
Audit job stuck in RUNNING	The Cloudflare Workflow is waiting on an external dependency (e.g., PageSpeed Insights API is slow or retrying).	Workflows have built-in retry and timeout behaviors. Wait up to 10 minutes. If the job does not complete, it will be marked as FAILED and you can retry.
Audit fails with URL_BLOCKED	The target URL did not pass Mincemeat's safety preflight checks.	Ensure the URL is public. Private address spaces, internal endpoints, and localhost references are blocked. See URL safety rules below.
Run fails with a verification signature error	The callback signature between the audit worker and coreapp is invalid or timed out.	Check that the system clocks are synchronized. If the issue persists, contact your Mincemeat administrator to verify the worker's shared secret configuration.
Audit finishes but shows a PARTIAL_SUCCESS status	Page body size exceeded the 2 MB limit, or the PageSpeed Insights API timed out for one device type.	Check the diagnostics section of the report. Mincemeat will still generate findings for the engines that completed, but no composite readiness score will be computed.

---

Target URL safety

To prevent abuse, security vulnerability scans, and denial-of-service attempts, the Site Audit system runs target preflight validations. A target URL is rejected if:

• It uses any protocol other than http:// or https://.
• It contains inline basic authentication credentials (e.g., https://user:pass@example.com).
• It targets ports other than 80 or 443.
• It resolves to a private IP range (RFC 1918, loopbacks, link-locals, or cloud provider metadata endpoints).
• It returns more than 5 redirect hops during resolve.

---

Robots.txt and crawler compliance

Mincemeat's edge crawler uses the following User-Agent:

MincemeatAuditBot/1.0 (+https://mincemeat.app/bots/audit)

• When auditing same-origin links, the crawler parses and respects robots.txt directives matching MincemeatAuditBot. If no specific rules exist, it falls back to the wildcard * directives.
• Disallowed paths in your robots.txt are skipped during crawling. These skipped pages will be logged in the links CSV artifact as BLOCKED_BY_ROBOTS.
• Metadata files like /robots.txt and /sitemap.xml are always fetched directly to analyze technical SEO, regardless of crawlers configuration.

---

Technical limitations

When planning your audits, keep these engine design parameters in mind:

• Single page focus: Mincemeat Site Audit analyzes only the explicitly entered target URL page. It does not perform deep multi-page recursive site crawls.
• Link crawl count limit: Same-origin link inspection halts after scanning 200 links to avoid worker timeout.
• No authentication support: The crawler cannot log in to pages behind basic auth, session cookies, or forms. Only publicly accessible pages can be audited.
• No browser rendering in Link Crawler: The link checker parses the initial HTML response. Dynamically rendered links added by client-side Javascript after loading are not caught.

Related

• Site Audit overview
• Use a correlation ID
• Contact support