PMM-Server is not reacting

[Matthias_Koster]

I installed PMM 1.0.4 a few days ago and I also tried it since 1.0.2 and I always run into the same issue.
I monitor 7 MySQL instances, versions from 5.6 to 5.7. When I install it and add the instances, everything works fine and i can use the metrics monitor, as I would like to. But after some days, the last time about 2 days, I cannot reach the metrics monitor website and also cannot login to the server itself. After I power the server off and turn it on again, everything works fine again.

The PMM-server is installed on a dedicated VMWare VM with Oracle Linux 7. The server has 4 CPUs and 6GB of RAM. The HDD is sized with 24 GB. I reduced the Prometheus metrics to 1day, so that it should fit onto the disk. I had the same issue with less instances, so i don’t think the disk is the problem. The attached image shows the output of df -h after the restart of the server.

Six of the monitored instances have aout 1.500 tables.

Could you give me any hint, what to do to solve this problem?

[weber]

Hmm, something to do with the virtual machine or docker itself it seems to me. So the machine hangs and you can’t check anything right?
Can you look at the system logs (/var/log/messages etc.) for potential issues? It could be even a host machine issue that operates with VMs which runs docker, heh.

[Matthias_Koster]

I can look at it, after the power off and on, but I didn’t find anything. I attach the last lines from yesterday in the evening, before the server did not respond anymore and the first lines after power on.