currently, I'm working on a new Plugin: Aggregate
Planned Features:
basics
- is based on existing graphs with existing data sources (=rrd files)
- will not create any new rrd file
- will create a new Graph based on a list or previously selected Graphs from Graph Management
- will not use Graph or Data Templates (is not related to a Graph Template, in other words: Graph Template = NONE)
- will prepend each GPRINT line with the target host's name
- will act as a normal graph, so it will allow for usual modifications
- optionally use "Aggregate Templates" to change graph items behaviour when aggregating them
- Aggregate Templates or to be defined manually. They lokk very much like graph items. Following features are supported:
- suppress a graph item (used e.g. to suppress 95th Percentile Comments that are of no use for an aggregated graph)
- override Graph Item Type e.g. to change an AREA for multiple traffic graphs to a STACK
- override CF type. Currently I have no real use for this. But it was easy to code ...
- override CDEF e.g. to show Traffic Output on the negative y-axis (cdef=current_datasource,-8,*)
- override sequence. Straighformat operation for e.g. a Graph with 2 data sources would be to have the aggregates items ordered first by host, then by data source (e.g. all items below the previous one). To have aggregates ordered first by data source (e.g. all input ds grouped together and all output ds grouped together), you have to change the sequence
- change color. In most cases you will aggregate graphs, where same colors occur on each basic graph. To override this, you may define a new color
But currently I'm not satisfied. I tried to explain this plugin to the participants of the 1.CCC.eu Cacti Community Conference. Apart from a coding error I had the impression, that these aggregate templates are too complicated.
Please see my alternative approach in the next post.
Reinhard