Hello,
I have some vms (and containers) which are intentionally stopped.
Every time they get backed up a pair of problem solved / started alerts is produced.
I understand this is because the backup routine starts the vm in a kind of paused state, then stops it.
I wonder what is the best way to solve the issue?
I have found a related post: https://www.zabbix.com/forum/zabbix-...p-vm-templates describing the process of filtering out useless alerts based on attributes of a VM.
My first idea was to add a filter to suppress warnings for VMs which do not have 'onboot':1 attribute in the config.
I have looked through https://git.zabbix.com/projects/ZBX/...p_proxmox.yaml, but it appears it does not pull VM config.
I suppose I could copy this section https://git.zabbix.com/projects/ZBX/...xmox.yaml#1427, adjust for pulling config, extract 'onboot' attribute.
On the second thought, perhaps it is better to change the alert logic and do not consider stopped VMs undergoing backup as really running?
I have noticed that during backup status is "running" but qmpstatus is "prelaunch".
A normally stopped state:
A normally stopped VM undergoing a backup:
A normally running VM does not see to change it status when backup is in progress:
If you ask me, Zabbix should better use 'qmpstatus' which, to my understanding, reflects the actual state of the emulator.
But I am not really familiar with VM lifecycle and I did not investigate behavior of backup in modes different from 'snapshot', so could use an advise...
I see that the template was authored by "Andrew Biba", but not sure how to get his attention.
I guess I could file a bug with Zabbix, but not sure if this qualifies as a bug...
Best regards,
Alex
I have some vms (and containers) which are intentionally stopped.
Every time they get backed up a pair of problem solved / started alerts is produced.
I understand this is because the backup routine starts the vm in a kind of paused state, then stops it.
I wonder what is the best way to solve the issue?
I have found a related post: https://www.zabbix.com/forum/zabbix-...p-vm-templates describing the process of filtering out useless alerts based on attributes of a VM.
My first idea was to add a filter to suppress warnings for VMs which do not have 'onboot':1 attribute in the config.
I have looked through https://git.zabbix.com/projects/ZBX/...p_proxmox.yaml, but it appears it does not pull VM config.
I suppose I could copy this section https://git.zabbix.com/projects/ZBX/...xmox.yaml#1427, adjust for pulling config, extract 'onboot' attribute.
On the second thought, perhaps it is better to change the alert logic and do not consider stopped VMs undergoing backup as really running?
I have noticed that during backup status is "running" but qmpstatus is "prelaunch".
A normally stopped state:
root@elon49:~# pvesh get /nodes/elon49/qemu/100/status/current --output-format json | jq . | grep status
"qmpstatus": "stopped",
"status": "stopped",
"qmpstatus": "stopped",
"status": "stopped",
root@elon49:~# pvesh get /nodes/elon49/qemu/100/status/current --output-format json | jq . | grep status
"qmpstatus": "prelaunch",
"status": "running",
"qmpstatus": "prelaunch",
"status": "running",
root@elon49:~# pvesh get /nodes/elon49/qemu/101/status/current --output-format json | jq . | grep status
"qmpstatus": "running",
"status": "running"
"qmpstatus": "running",
"status": "running"
But I am not really familiar with VM lifecycle and I did not investigate behavior of backup in modes different from 'snapshot', so could use an advise...
I see that the template was authored by "Andrew Biba", but not sure how to get his attention.
I guess I could file a bug with Zabbix, but not sure if this qualifies as a bug...
Best regards,
Alex