Cross-region health checks were silently 404ing in production

Spent a chunk of yesterday debugging why FlareWarden's "Is Website Down?" tool wasn't returning cross-region results in production. Worked perfectly locally. The tool checks a URL from multiple Fly.io regions to give you a global picture, and the multi-region part was just... not working.

Turned out to be a host routing issue. When one Fly machine pings another internally, the Host header is an internal IP, not a real domain. My HostRouter didn't recognize it, so the request fell through to the status page router, which returned 404. Three lines changed in the middleware. Classic "obvious in hindsight" bug that took way too long to find because the logs just showed clean 404s with no context about why.

While I was in that area of the code I also added dynamic OG images — so when someone checks ?url=https://example.com, the social preview shows "Is example.com Down?" instead of the generic title. Small thing but it makes shared links look way more intentional.

The rest of the day was SEO gruntwork. Added FAQ accordion sections and JSON-LD structured data across all 19 marketing pages. Not glamorous but Google eats that stuff up. Also fixed a CSS issue where decorative SVG underlines on headings were stretching to full container width on mobile instead of hugging the text. Had to add   between the last two words on 26 headings to prevent orphaned words at mobile breakpoints too. Touched 21 templates for that one.

Some days you ship features, some days you fix underlines on 26 headings.