BMC Health

[Andrew Jeffery]

On Fri, 23 Mar 2018, at 10:29, Patrick Venture wrote:
> On Thu, Mar 22, 2018 at 4:44 PM, Andrew Jeffery <andrew at aj.id.au> wrote:
> > On Fri, 23 Mar 2018, at 08:33, Patrick Venture wrote:
> >> On Thu, Mar 22, 2018 at 1:10 PM, Andrew Jeffery <andrew at aj.id.au> wrote:
> >> > On Fri, 23 Mar 2018, at 05:24, Patrick Venture wrote:
> >> >> The Facebook presentation yesterday talked about BMC health, free
> >> >> memory, etc.  I didn't see anything in the IPMI spec about this
> >> >> information, and beyond my general drive for metrics, just system
> >> >> health would be great.
> >> >>
> >> >> Things I care about:
> >> >> up-time
> >> >> free memory
> >> >> ... other things?
> >> >
> >> > * Free flash space
> >> > * Network traffic stats
> >>
> >> I actually have an OEM IPMI command for grabbing the ethernet
> >> statistics that loads as a library into phosphor-host-ipmid.
> >
> > Ok. I don't know how best to export the metrics, though we (others inside IBM) are already looking at the problem.
> >
> >>
> >> > * CPU load average
> >>
> >> Basically a cleaned uptime result?
> >
> > I guess? I don't really mind how we get the data, or how it's exported, just that if we're going to do metrics I think these values should be included.
> 
> Right.  I'm reviewing this from an IPMI perspective, things that can
> be reported that way, instead of a network interface -- the network
> interface being more .... solved as there are plenty of packages for
> exporting metrics over the network.

Looks I should have absorbed more of your original mail :)

Cheers,

Andrew