[rrd-users] One-off and no-data-problem

[Simon Hobson]

Alex van den Bogaerdt wrote:
>  > It was, but it's not enough for rrd tool to calculate a value for the
>>  whole period.It knows that at time 400 the value was 123. Before that
>>  it has to assume the values are unknown as you've not supplied any
>>  earlier updates.
>
>To avoid confusion: assuming values are unknown before 300000400 is wrong.
>  After all, rrdtool updates are done at the end of an interval, not at the
>beginning.
>
>For COUNTER:
>at time xyz your counter value is x
>at time xyz+300 your counter value is x+y
>rrdtool computes the rate at time xyz+300. The resulting rate is valid
>between times xyz and xyz+300.
>After this, normalization is done to make a PDP, and consolidation to get
>a CDP.
>
>GAUGE is no different. The only difference at this stage is that it needs
>not do (x+y)-x and (xyz+300-xyz). The "resulting rate" (the number being
>fed to RRDtool) is still valid bewteen xyz and xyz+300.

How does it know how far to go back when there has been no prior update ?

For a counter type, there is no prior version, and so it is 
impossible to calculate a rate - unless you assume a value (zero ?) 
which IMO would be worse than assuming unknown.

I know you can apply the update value to the preceeding time, but how 
far back ?  To the beginning of the step period ? To the beginning of 
the RRD (start time in the create command) ? As far back as the 
heartbeat time ?

Genuine question, and similar questions arise if no updates are 
received for the heartbeat interval.

-- 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.