Slow NPM performance

[dan-burt]

On my Unraid server, I have always run NPM as a reverse proxy pointing to other Docker containers. This past week, following a system reboot (though I don't know if, how or what this changed!) the performance of NPM is very bad, with pages taking a long time to load, others showing a holding "disconnected" page.

If I look into the error logs, all of the proxy hosts have this recurring entry:

upstream server temporarily disabled while connecting to upstream

In trying to resolve these issues, I have also upgraded the docker container version of NPM, but also the other application containers.

I have only created the Nginx config through the NPM UI, so shouldn't expect any issue. In all of my use cases, the other containers are running over HTTP (not S). NPM provides the SSL wrapper. I use the Unraid server's hostname ("tower"), its IP address (192.168.1.1) or localhost (127.0.0.1) in the NPM UI.

If I access the container directly using it's local IP address, performance is as expected.

[dan-burt]

Issue was related to local LAN IP addresses & DNS getting messed up. Saw incorrect upstream IP addresses being used in the NPM error logs.