[BUG] Some of the graphite metrics are broken in the latest PVE

[almku]

I'm not quite sure this issue applies only to the very latest version. The following is seen in the graphite relay log

Spoiler: carbon-c-relay log

[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata0.timed_stats ARRAY(0x3d1ee58) 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata0.account_failed true 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata0.account_invalid true 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata1.timed_stats ARRAY(0x3d0d970) 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata1.account_failed true 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.sata1.account_invalid true 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.blockstat.ide2.account_invalid true 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.status running 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.name mcc 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.template 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.1000.qmpstatus running 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.200.template 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.200.qmpstatus stopped 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.200.name pf 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.qemu.200.status stopped 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.lxc.101.status running 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.lxc.101.lock 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.lxc.101.name log 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.lxc.101.type lxc 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.lxc.101.template 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.local.content iso,vztmpl,backup 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.local.type dir 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.local-lvm.content images,rootdir 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.local-lvm.type lvm 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.ssd225.content images,rootdir 1489650955\n"
[2017-03-16 07:55:55] I bad message: "pve.storages.pve1.ssd225.type lvm 1489650955\n"

 

[dcsapak]

afaics those values are all non-numeric and graphite supports only numeric values, so this is normal

there were no groundbreaking changes in the graphite export, so i guess this was the behaviour all along
maybe someone finds time to clean it up to only send numeric data

 

[Daniele de Petris]

I have the same problem proxmox send dirty data to my graphite collector, this problem is fixed on proxmox 5 ?