I'm seeing a strange behaviour happening on publisher machines, after they are online for a couple of days.
After working normally for a couple of days (2-3), without any warning or visible issue, we stop getting access to specific resource on publisher console like, host:port/miscadmin or host:port/system/console/configMgr
If we try to hit one of those URLs we got a 504. We are still able to use and login on host:port/crx/packmgr/index.jsp for instance, and we are able to load pages and content from this publisher instance as well.
I couldn't find any issue on CPU or memory, instances are not using any swap, and have free memory, and CPU never goes over 20% on this instance.
error.log don't have any logs when we trigger the 504 URLs.
At last but not least, the issue disappears after a restart for a couple of days/hours. And on that, the instance wouldn't stop running the usual /crx-quickstart/bin/stop command, it needs to be killed.
I'm using Adobe Experience Manager (184.108.40.206).
Can you provide some ideas on how to troubleshoot this issue or how can we solve it?
Which version of Java you are using to run AEM? It is JDK 11 or JDK 1.8?
Also do you get any blank page when you access the content pages? Is this issue only on Publish instance or on Author as well?
We are using JDK 1.8
[XXX]$ java -version
java version "1.8.0_251"
Java(TM) SE Runtime Environment (build 1.8.0_251-b08)
Java HotSpot(TM) 64-Bit Server VM (build 25.251-b08, mixed mode)
We get the content pages as normal, they are working without issues.
Only detected in Publish instance.