Connection Reset By Peer at Exactly Sixty Seconds, Every Time

Every failure logged at the same number of seconds of idle time. That regularity is the diagnosis: a load balancer idle timeout killing connections your…

Connection Reset By Peer at Exactly Sixty Seconds, Every Time. Abstract error autopsy illustration in orange and dark grey on debugly.dev
Middlebox Typical default idle timeout
AWS ALB / NLB 60s (ALB), 350s (NLB)
GCP HTTP(S) LB 600s
Azure LB 240s (about 4 min)
Linux conntrack / NAT often 120s or configurable
nginx keepalive_timeout 65s server side

Published on debugly.dev by Rohit Bhadani | Founder of Krova Cloud