JRDS can managed quite an high number of nodes, finding it's way within a long lists of nodes can become difficult.
So jrds provides a way to create tree using only a subset of possible graphs, using filter.
A filter is defined by a xml document, as usual, giving
- it's name
- somes tags, only hosts that matches all the tags will be displayed
- some regex filtering graph's tree, any graph whose path match a regex will be displayed.
The doc type for a filter is
<!DOCTYPE filter PUBLIC "-//jrds//DTD Filter//EN" "urn:jrds:filter">
And the DTD is at github
All defined filters are show in the tab “All filters”
So for a graph to be displayed, it must met two conditions :
- all of the given tags must be in the hosts tag
- it's path must match one of the given path.
For example, given the following filter configuration :
<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE filter PUBLIC "-//jrds//DTD Filter//EN" "urn:jrds:filter"> <filter> <name>x86 calculation servers</name> <tag>calc node</tag> <tag>x86</tag> <path>/Sorted by view/System/Memory/.*</path> <path>/Sorted by view/.*cpuraw.*</path> </filter>
A tree called “x86 calculation servers” will be created, and it will display the memory usage and cpu usage for all servers that are tagged by both
calc node and