Re: pmie rule for instance disappearing

[Jonathan Knispel <jkwaoz@xxxxxxx>]

On Friday 02 May 2008 13:14:13 Michael Newton wrote:
> we need a pmie rule to notice when any instance on a metric disappears.

> but the problem is that actually if the instance disppears,
> *both* cluster.blah *and* cluster.blah @1 become undefined.

  I'm pretty rusty but that sounds like a bug.  A rule that uses @4 and @5 
probably shouldn't stop "working" until time 4 * delta after the metric first 
disappears.  [ Down with historical revisionism!  ;) ]

> while im about it, wouldnt u expect count_inst( cluster.blah >=0 )
> to evaluate zero when there are no instances? and sum_inst ?

  Seems very reasonable.

  Having said those things, I do vaguely recall there were some fairly subtle 
issues dealing with zeroes and undefined values back in the early days.

Regards,
Jonathan