Nimbostratus
MVPNot an expert on optimization but I remember the videos about optimizing Nginx. Also maybe as you are using Nginx as reverse proxy you could need to take into account clientside and serverside connections at the same time.
Performance-Tuning NGINX Open Source and NGINX Plus | F5
Tuning NGINX for Performance (f5.com)
Class 8: Performance Tuning NGINX Plus (f5.com)
Maybe to optimize the server side you could see TCP multiplexing (like F5 OneConnect) profile.
Load Balancing with NGINX and NGINX Plus, Part 2 (f5.com)
NimbostratusHello, Nikoolayy1
Thanks for the reply. But my main question is how to predict that the situation is getting worse.
MVPkoef you can send the logs to a external SIEM like ELK or splunk and have dashboards that show the connections called metrics and see if they start to increase over the days or weeks. If you are in openshift/kubernetes prometheus is an option as well.
Nginx | Documentation (elastic.co)
GitHub - nginxinc/nginx-prometheus-exporter: NGINX Prometheus Exporter for NGINX and NGINX Plus
Also you can review the new Nginx One Cloud GUI dashboard :
Product Preview: NGINX One | F5
F5 NGINX One Cloud Console | NGINX Documentation
This Relic stuff also seems nice but I have not used it:
Monitoring NGINX and NGINX Plus with the New Relic Plug-In | NGINX Documentation