No spike in CPU. Error budgets looked fine at a glance.
Users still reported blank pages in a thin slice of traffic.
Logs only showed upstream resets with no clear application exception.
The culprit was a stale keep-alive timeout between nginx and the app.
Aligning idle timeouts stopped the intermittent 502s within an hour.
We also added a dashboard for upstream reset reasons so the next page is faster.
Totals · 1 entries · 0 followers
Newest first
by Julia Berg
No spike in CPU. Error budgets looked fine at a glance.
Users still reported blank pages in a thin slice of traffic.
Logs only showed upstream resets with no clear application exception.
The culprit was a stale keep-alive timeout between nginx and the app.
Aligning idle timeouts stopped the intermittent 502s within an hour.
We also added a dashboard for upstream reset reasons so the next page is faster.
Separate opinion from lived experience. Mark what you have actually done.
Sign in to verify experiences. Authors cannot verify their own experiences.