Just an update on here if it helps anyone in the future…
I think I may have got to the bottom of it but it’s still early days and there’s still some unanswered questions.
In Zabbix, it showed me that the threads are constantly climbing. Where you see the plateau on the below image is where I see the issues and the JVM is not able to create any more threads.
However, upon further research, I came to the conclusion that the number of threads is only limited by RAM, Heap size and CPU. Given that my RAM is only 20% used, my Heap size is only 15% used and my CPU avg is 4%, I didn’t think this would be a cause. My Java knowledge is rather sketchy too so I thought this may possibly be expected.
However, in the absence of any other avenue of investigation, I thought I’d look into the threads and what they’re doing and why they’re climbing so I used a JVM profiler to get visibility of the threads. When I came back in the morning, there were hundreds of parked threads with the name
:coapEndpoint-UDP-224.0.1.187:5683
Then I remembered an error that I have intermittently seen in the logs since the 4.2 OH upgrade
Stopped without executing or closing null
which a previous search told me is from the Shelly Binding but as no one else reported any major issues, I didn’t think too much of it.
So, I deleted the shelly Binding (and switched all my Shelly devices to MQTT) and low and behold, the thread count is now totally stable.
Like I say, it’s early days and I can’t say yet that the issue is fixed. I saw memory issues every 3-4 days but this is the impact on the thread count now that the Shelly Binding has been removed:
So it looks like the Shelly Binding is not disposing of threads correctly.
@markus7017 FYI
EDIT: Just spotted OpenHab 4.2.1 stopping after 3 days - OutOfMemoryError. That user also has the Shelly Binding installed.