Audit with 32 URLs times out, audits splitting them are successful

[kokimo1]

Describe the bug
Hi, we have an audit named "UIC - SDD - Red template tracking" (at https://app.equalify.uic.edu/audits/3174659f80c84148af89f4b220fec47d) that we're using as a baseline to track & triage template-level a11y issues in the Red Multisite codebase (32 urls from different sites). I think content on one of the sites (or combination of sites) is causing the scan timeouts we've been seeing lately. I initially thought the timeouts were related to Red's hosting issues, but that was resolved earlier this week.

A few things to note:

• This audit previously ran without issue. It initially flagged 245 blockers on July 28, 2026 but has failed since then
• Individually, it looks like each of the 32 URLs can be scanned successfully in a Quick Scan.
• I created two new audits splitting the URLs between them and they both scan without issue:
  • UIC - SDD - Red template tracking (1st half)
  • UIC - SDD - Red template tracking (2nd half)
• I did a test in headless Chrome (Puppeteer), and it successfully loaded all 32 uls in sequence. Took about 3 minutes.

To Reproduce

1. Go to UIC - SDD - Red template tracking. This is the one with 32 URLs.
2. Click "Scan Now"
3. The scan completes with "Scan Timed Out" warning, "Scan Errors (1)", and zero blockers.
4. Scan error detail says "Scan timed out after 15 minutes of inactivity"

Expected behavior
The 32-URL audit should complete successfully and report blockers, as it did on July 28, 2026 (245 blockers).

Screenshots

Additional context

• One URL in the audit (https://red.uic.edu) has a broken third-party iframe that increases page load by about a minute. However, this URL also scans successfully as an individual quick scan, and the other 31 URLs all load in under 3 seconds in a real browser.
• The 0-blocker result from timed-out scans overwrites the previous valid baseline, this is tangentially related to issue Failed scan nukes existing blocker IDs #654.