502 Error with Nginx

502 errors are typically due to issues at the web server or proxy level rather than at the Appian/Jboss level. See - www.checkupdown.com/.../E502.html

That said, try bypassing the webserver and access the application server directly:

1. Stop JBoss and change SERVER_AND_PORT in custom.properties to be the <JBOSS hostname>:8080
2. Start JBoss and access the reports, admin console etc. that were previously showing errors. Make sure to use the URL http://<JBOSS hostname>:8080/suite/tempo etc.

If you are using web servers like Apache or IIS then make sure your web server is pointing to your Appian16.2 directory.
For Apache, httpd.conf & appian.config files has to be updated to point the web server to new code base.

Abhi - after making the changes you suggested, we are still able to access all components locally as expected, but we are still experiencing the same 502 issues with certain records, reports, etc when accessing publicly. We agree that the issue likely stems from the web server, but cannot identify the cause. We are using nginx and JBoss.

If you're bypassing the web server (using Jboss hostname and port 8080) and see the same issue then the web server wouldn't come into play anymore. Can you attach the latest server.log please? I would also suggest clearing out the browser cache completely and test again

Is it possible to bypass nginx?

Hi Abhi - I'm working along with Ashvin on the above issue. Please see the attached for the latest server.log file. In addition, I've added a screenshot of the error message that I'm seeing.

In this example, I'm trying to access the OOB Appian 'Users' record. An interesting note though is that given certain search parameters, the record will load without issue. In the second screenshot, I've added '/search/asdf' to the url and now the error message is gone.

Unfortunately, clearing browser cache and JBoss osgi-cache did not resolve it. Lastly, bypassing nginx won't be possible as we are using it as a router to multiple different VMs on the machine.

server.log

Do the logs on nginx show anything wrt the root cause of the 502s? This looks to be fairly common. See - asdqwe.net/.../

Hi Abhi - Thanks for the suggestion, you are definitely pointing in the right direction. I was able to find the nginx error logs and it told me that "upstream sent too big header". After a few google searches, I was directed to stackoverflow.com/.../upstream-too-big-nginx-codeigniter . After adding the proxy_buffer_size, proxy_buffers, and proxy_busy_buffers_size settings to the nginx.conf file, all records and reports were able to load as expected! It looks like we are good for now, and thank you for your help!