Why the first crawl may not open an incident
Understand baseline-only observations and current-state exceptions.
Last checked 21 August 2026
Many TechDash checks look for a change, not just a current value. A change cannot be established until TechDash has two usable observations. For those checks, the first completed check records a baseline and returns a pass instead of opening an incident. A pass appears as Good, does not reduce Health Score, and does not send a failure notification.
What the first crawl records
Depending on which checks can run for the page, the baseline can include:
- Title and H1 text
- Visible body content and its fingerprint
- Canonical and indexability state
- Structured data
- Analytics identifiers
- SEO response headers
- hreflang values
- Internal and external link signals
- Product, video, mobile, image, and WebMCP signals
For some groups, TechDash also checks whether that specific check has run before. This prevents a newly introduced or newly enabled optional check from treating existing website content as a fresh regression merely because no earlier result exists.
You may see messages such as Baseline stored., Title baseline stored., or Canonical baseline stored. under Latest results. These are expected healthy, passing outcomes.
What can still open on the first crawl
Not every check needs a historical comparison. Current-state checks can fail immediately when the first crawl supplies enough evidence. Examples include:
- An HTTP error or fetch failure
- A first-hop redirect
- A page that looks like a soft 404
- An invalid current structured-data state
- A current site-resource failure
- An invalid TLS certificate
- A DNS resolution failure
- A certificate or domain already close enough to expiry to meet a threshold
The exact result depends on the check. For example, change-oriented noindex checks compare the new state with a prior state, so an existing directive may establish the starting condition rather than be described as newly added.
There is also a difference between opening an incident and sending an alert. A failing latest result can appear in Incidents immediately and affect Health Score according to its Advisory, Warning, or Critical severity. Page-check desktop and webhook alerts require a change from pass to fail. An unknown first state that fails immediately does not produce that change alert.
Website setup redirects
During initial website enrichment, TechDash follows the homepage and can adopt a valid landing hostname. It updates the saved website hostname and origin when the landing hostname is valid and not already represented by another website. This setup normalization is separate from a later monitored-page Page redirects incident.
If the landing hostname would duplicate an existing website, TechDash removes the duplicate setup entry and shows a warning instead of silently merging the two.
If you expected an incident
- Open Websites and select the website.
- Open Monitored Pages.
- Expand the relevant page and review Latest results.
- Look for a baseline, pass, skipped, or error result for the check you expected.
- Confirm the page and the individual optional check are enabled.
- Confirm the response is HTML if the check requires rendered page content.
A skipped result is not a hidden incident. An earlier required check failing, a non-HTML response, a disabled optional check, missing evidence, or the browser being unable to finish loading and inspecting the page can prevent dependent checks from running.
Allow a later successful crawl to establish the comparison. You can use Run now from the monitored page when you need another full page run. Use Check & Verify only when an incident already exists and you want to rerun the work relevant to that incident.