Endpoints whose last saved audit was hard down (no response), a 5xx server error or very slow,
plus every endpoint with an open trouble ticket. Re-audit checks just these again from this browser (one GET each, like Run Audit),
then you can save the new results, open or update tickets for the ones still failing, and resolve the tickets of the ones that recovered.
Save results writes Last Audit Date, LastStatusCode and LastResponseMs on the re-checked endpoints (what Performance and this page read);
it doesn't add an audit to History. Tickets follow the problem types ticked for tickets in config.js → tickets.problems, and you confirm each step first.
Endpoints on hosts with a live host exception are listed but can't be re-audited. Times are measured from this browser.