hello
Perhaps I didn’t explain this well.
You have two heap settings, One for Elasticsearch and One for Graylog (i.e JAVA) So in this case Your logs are complain about memory.
docker-compose up
Which show the info above, normally when troubleshooting issue/s we “tail log files” in this case using docker logs -f maybe help find something more we missed, just a thougt.
The thing is I just upgrade mine when you had this issue, I didn’t not have a problem with Graylog starting up ,nor did I have any information with lack of memory. The suggestions I offer is for troubleshooting and perhaps find something that was missed.
Understood, I was suggesting to NOT stop those container and leave them running since they are running fine, I was suggesting ONLY stopping the Graylog container replace the image and starting it back up, just for troubleshoot.
What I did notice in the compose file were changes in the configuration from default to some custom settings, not sure if that would make a impact. I also realize this was running fine before, maybe the new image does like custom settings, just a thought
EDIT: Did some digging around the web.
In your logs it shows a few of these errors before it fell on its face.
[warning][os,thread] Failed to start thread “GC Thread#0” - pthread_create failed (EPERM) for attributes: stacksize: 1024k, guardsize: 4k, detached
Came across couple old post but had the same issue.
There were a couple solution in those not sure if they pertain your issue.