Re: [pcp] RFC: derived metrics in PCP archives

[Ken McDonell <kenj@xxxxxxxxxxxxxxxx>]

On 04/12/15 10:19, Nathan Scott wrote:
> ...
> > > So we end up with a PMID 511.*.* in the archive's metadata and in the
> > > archive's pmResults ... kaboom!  libpcp gets very confused trying to
> > > handle this because it looks like a derived metric but behaves like a real
> > > metric in the archive.
> Ouch.  :)
Not really ... just outside the scope of the original design ... 8^)

> ...
> A couple of alternatives - there are some free ranges earlier in the stdpmid
> set, which we've begun reusing (24 would be the next - grep on "FREE SLOTS").
> Another option might be to reuse the PMI_DOMAIN (245) for these metrics too?
24 is just as good as 384 ... keeping our footprint compact has some 
advantages.  I'd like to avoid PMI_DOMAIN and indeed any <domain> that 
could be used by any PMDA (the importers are like a PMDA in this context).
> ...

Thanks Nathan.

I think my suggestion is still in play ... I'll wait a few more days for 
any other comments, then I'll likely just "do it".