upstream connect error or disconnect/reset before headers. retried and the latest reset reason: conn

[SergiSeira]

Hello,

we are getting this message very, often while browsing our project. If you reload the page it works fine for a while until another of these messages pops up.

Our project has a load balancer to an autoscaling instance group. There're no errors anywhere in the healthchecks o in the group itself.

Logs in the load balancer show 503 errors with message :

 failed_to_connect_to_backend

And when failing a curl -I shows :

HTTP/2 503
content-length: 114
content-type: text/plain
date: Wed, 15 Feb 2023 09:38:31 GMT
via: 1.1 google
alt-svc: h3=":443"; ma=2592000,h3-29=":443"; ma=2592000

Anyone experiencing these problems?

Thanks!

[sijohnmathew]

Hey

We are also facing the same issue. Did you get any response ?

Sijohn

[SergiSeira]

Hello,

we got no response ... we had to point our dns records to a single big instance of the group, which luckily our system allowed and waited 24-48 hours to try again with the loadbalancer and instance group if the issue disappeared, and it did. 

We yet don't know what was the issue.

Regards