Question
Under a traffic surge (a marketing email went out 10 minutes ago), your reverse proxy starts returning 502 Bad Gateway for ~6% of requests, climbing. The proxy's own CPU is moderate (~60%). Upstream app pods show no crashes and their own request latency is normal for the requests that succeed, but their accept queue / connection backlog is growing and some new connections to them are being refused or slow to accept. The proxy logs show 'upstream prematurely closed connection' and connect timeouts to upstreams. Autoscaling on the app tier hasn't added pods yet. Triage and mitigate.
Stop the bleeding first (mitigate), then form hypotheses from real signals. Separate root cause from symptom, communicate status as you go, and close with what prevents a repeat.