how to monitor disk I/O with zabbix 1.9.x

[vnline]

I have search so long about zabbix monitor disk I/O ,but I can't find anything helpful.anyone can tell me how to monitor disk I/O with zabbix 1.9.x?

sorry for my poor english!

[nelsonab]

vnline, what are you trying to accomplish? Do the vfs.* items not provide what you need? http://www.zabbix.com/documentation/...l/config/items

[vnline]

vnline, what are you trying to accomplish? Do the vfs.* items not provide what you need? http://www.zabbix.com/documentation/...l/config/items

first,thanks for you reply~

I known the vfs.* items can provide what i need.but I can't find offcial documentation guide how to setup.

[vnline]

I get it.things a lot!

[CeeEss]

vnline, what are you trying to accomplish? Do the vfs.* items not provide what you need? http://www.zabbix.com/documentation/...l/config/items

barging in here ... pardon me

Thanks for the pointer to vfs.dev.[read][write]. T'is good they return actual numbers, but the values that are returned are allegedly avg[1][5][15], so i'm guessing the agent side is using something like sar? Q: how (exactly) are these agent params obtaining and calculating disk data averages? I've compared sar and iostat output against data returned by agent and they don't correlate. Is there a relationship between agent polling/reporting frequency and the averaging period of returned data?

The vfs.dev. items weren't included in the Linux or BSD templates nor assocated with any template in data.sql for 1.8.6 or 1.8.10 (they get put into a "help_items" table - whatever that's for) . Are they ready for primetime?

thanks!

[CeeEss]

Got it. This works really well! Took a while to realise that the agent is performing time averaging and a delta calc isn't required in the item definition.

barging in here ... pardon me

Thanks for the pointer to vfs.dev.[read][write]. T'is good they return actual numbers, but the values that are returned are allegedly avg[1][5][15], so i'm guessing the agent side is using something like sar? Q: how (exactly) are these agent params obtaining and calculating disk data averages? I've compared sar and iostat output against data returned by agent and they don't correlate. Is there a relationship between agent polling/reporting frequency and the averaging period of returned data?

The vfs.dev. items weren't included in the Linux or BSD templates nor assocated with any template in data.sql for 1.8.6 or 1.8.10 (they get put into a "help_items" table - whatever that's for) . Are they ready for primetime?

thanks!