Private Cloud NGINX Results in 502

Hey everyone! Reaching out with a hope that someone out there has experienced this issue and may be able to offer some guidance/wisdom. We’re using NGINX as a middle layer between client and server. Our users were reporting 502 errors at random points during app use. These 502s were occurring mainly during /xas calls. We found out that NGINX has a maximum number of keepalive requests that can be made using a single keepalive connection before the connection would be forcibly closed. Generally this would be fine, but the issue we’re experiencing is that NGINX is not releasing/reusing existing requests. So eventually a user reaches the limit just by working in the application and their connection is forcibly closed by NGINX. At this point the user needs to create a new connection in order to continue working in the application.   Has anyone else ran into this problem? If so, did you ever find a fix to get NGINX to release/reuse some of these requests? Thanks in advance!
0 answers