I'm afraid but I think you've got lost.
First, please understand what of those parameters provided by cacti's rra definitions are pur rra parms and what are cacti-only ones:
timespan is a parameter used for cacti only, all other parms are pure rrdtool ones.
Second, understand how cacti works with those rra definitions: When setting up a data template, you'll have to associate ALL rra's you want to be created for the file. But they must be in ascending order with respect to "consolidation". That is, the "first" rra defines the basis (pure data without consolidation". The second one defines the first consolidation level. The
steps give the number of primary data points to be consolidated. So it does not make sense to define another rra with the same interval size (5 min) twice.
Third, understand the legend, e.g.
Monthly (2 hours average). This is pure text. It's your task to change the text so it fits to the
timespan set for this rra. And, with respect to the famous
zoom feature, this text has merely lost it's meaning. It's only helpful for the initial display to know about this rra's consolidation level. Now, let's see the solution:
Last, understand rrdtool's
best fit algorithm:
Code:
The data fetched from the RRA is then consolidated so that there is exactly one datapoint per pixel in the graph.
If you do not take care yourself, RRDtool will expand the range slightly if necessary. Note, in that case the first
and/or last pixel may very well become unknown!
This says, that even if you have 1000s of data points (e.g. one months worth of 5 min intervals), there are only 500 pixels to graph them! So rrdtool will even
consolidate 5 min data points until the above (one consolidated data point per pixel) is reached.
More precise data is graphed only when zooming in!
Quote:
Time range
[-s│--start time] [-e│--end time] [-S│--step seconds]
The start and end of the time series you would like to display, and which RRA the data should come from.
Defaults are: 1 day ago until now, with the best possible resolution. Start and end can be specified in several
formats, see AT-STYLE TIME SPECIFICATION and rrdgraph_examples. By default, rrdtool graph calculates the width
of one pixel in the time domain and tries to get data from an RRA with that resolution. With the step option
you can alter this behaviour. If you want rrdtool graph to get data at a one-hour resolution from the RRD, set
step to 3.600. Note: a step smaller than one pixel will silently be ignored.
So, rrdtool always uses "best fit" rra's.
To sum up: Blow up the
first rra to the amount of data points needed. Keep your 4 rra's, no need for a fifth (unless you want to see five graphs when clicking on the default graph). And use the zoom feature to actually see the precise data
Reinhard