Rafal Gwizdala wrote: >- How do you calculate monthly statistics? That is, how do you deal >with months having different numbers of days?
I don't. I deal with simplistic approximations instead. On my graphs I show about 7 week (because that's what fits neatly in the number of pixels, and show total bandwidth used for the previous 30 days (on a rolling basis). If you want to deal in 'real' months then you'll need to do some work outside of RRD. It may involve calling RRD multiple times to get data from different timescales. For example, if you want to plot total events by calendar month for the last year, then you'll need to do 12 queries to extract the data for each month, and then plot those values (probably better done outside of RRD). In fact, for that sort of usage, you may actually find a different storage system to be more suitable. >- Does it mean the rrd file needs to start at the boundary of the >largest consolidation period, in your case at the beginning of a >year? This way your consolidation period is aligned to calendar and >spans exactly a calendar No, all periods (both primary and consolidated data points) have boundaries which are an integer multiple of the period (step size or step size * steps/CDP) since unix epoch (midnight, 1st Jan, 1970, UTC). Thus a consolidation of 1 day will always start at midnight UTC regardless of your local timezone. So for me at the moment, my "days" start at 1am local as we're on summer time. The start time in a database only determines the earliest timestamp of updates that may be used. Once you start updating, then the 'start' of the database will move - there are simply N buckets of T time (both determined when you create the database) and the 'end' of the database is simply the timestamp of the last update. You're oldest data (ie 'start' of the database) will simply be that time less the cumulative length of all the buckets - and it is different for each consolidation. -- 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. _______________________________________________ rrd-users mailing list rrd-users@lists.oetiker.ch https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
