Red Hat Bugzilla – Bug 1028084
RFE: Extended availability metric
Last modified: 2015-11-01 19:45:51 EST
Description of problem:
Currently, availability metric has for different resources different meaning.
started/stopped for AS7 Managed Server resource
started/stopped or no long existing for AS Server resource
present/missing for most AS7 subsystems, deployments, datasources
To describe resource state, it's not enough. For example datasource can be enabled/disabled (I don't mean at JON level). It depends on plugin developer, how he understands resource state. He can export this "enabled/disabled" either using Trait metric or as a configuration property.
In any case, for user it is hard to know the real meaning of availability. for example:
JON is reporting AS7 managed server as DOWN, this means:
AS7 managed server is stopped
or AS7 managed server was removed from HC outside JON
User has to find out probably by trying to start it, which would result in error if server was removed.
As a solution we should somehow extend availability metric to be able to describe such situations. Or maybe, just let plugin somehow bind trait with availability. For example:
plugin API would define types for extended availability, say : ENABLED, DISABLED
in plugin-xml, developer would set trait metric with some "availability" attribute which would bind collecting this metric with availability scan. Then in plugin XML, developer would have to map possible trait values to ENABLED or DISABLED.
Then in UI we could instantly show different availability icon based on actual availability and this extended one.
Version-Release number of selected component (if applicable):