I have started getting memory low errors from New Relic ever since I upgraded to the newest beta. When I checked the graph, it looks like this (the drop is where I last restarted the server):
What version are your running? Made any modifications to the code?
May have tracked it down. Noticed this pattern in New Relic:
So I checked the logs and found this:
I had briefly turned on ‘reply by email’ before deciding against it and disabling it and deleting the associated e-mail account. It must still be active though, and this is generating a number of errors. Going to see if I can get this to stop and see if that’s the cause.
does the issue go away if you reboot?
You’ll want to turn off the site setting, that’s how to turn it off.
No, I’ve tried twice now. I restarted again after posting the original post today, and since then I’ve noticed the slow crawl upward again at the same rate, even after shutting off all things related to reply-by-email:
This time, the rsyslogd resources are flat, but I’m seeing it creep upward with nginx:
I believe I have, and all the errors have stopped.
Can you get some more details, what process in particular is creeping up ?
The last time I checked the processes before doing a restart it seemed to be various unicorn workers that were hogging most of the RAM. I’m not a Ruby programmer though so I’m not sure if that’s useful to you or not - is there anything specific I can provide you that would be helpful?
It looks like the memory leak is somewhere else, there are multiple reports. @sam is investigating.