Fixed a false "everything is down" reading on our new status page

Our brand-new status page was reporting most of the platform as down when it wasn't. A single known, expected failure (like one site blocking automated checks) was dragging an entire category to "Major Outage" even when the vast majority of it was fine.

Shortly after launching our public status page, it started reporting 4 of 5 components as down. That wasn't true. The bug: several categories check dozens of individual links at once, and a single expected, known failure (one dead job listing, one community site that blocks automated traffic) was enough to mark that whole category "Major Outage," even when 90%+ of it was genuinely fine. It now reports status based on the real proportion of what's actually failing, not the single worst result.

We also found and are actively working through a real, separate issue: our mentor-profile sampling check has been failing to reach our backend consistently. We've added retry logic (the same pattern already used elsewhere in the app for this exact class of problem), and we're continuing to verify whether that fully resolves it in production. We'd rather tell you it's still being worked on than mark it fixed before we're sure.

← All updates