1. Open the monitor
From the dashboard, open the http monitor and scroll to What came back. Each entry is collapsed: the status line, when it was captured and which region saw it. Expand one to read the headers and body, or use Copy to paste it into a ticket.
2. Know what is captured, and when
Two snapshots per outage per region: the first failing result and the result that recovered. A continuing outage stores nothing more. Connection failures and timeouts have nothing that came back, so they carry no snapshot. Snapshots are also shown on the incident they belong to.
3. Know what is never stored
Headers named like credentials (authorization, cookie, set-cookie, x-api-key and similar) are stored as [redacted]. Values anywhere that look like tokens, API keys, card numbers or long hashes are stored as [scrubbed]. Only the first 16 KB of the body is kept. Everything is deleted after 7 days. Private monitors run by the agent never capture a snapshot.
4. Turn it off for a check, or on for your plan
On the monitor page, untick Keep a response snapshot and save to opt that check out. Response snapshots are included on Growth and above; on Free the monitor page shows how to enable them.
Go deeper
The full reference lives in the docs: Checks documentation. Error codes named above are each explained in the error-code reference.