I have two alert definitions: to-be-recovered:= if { list of conditions }, disable until manually enabled by recovery alert or manually re-enabled recovery alert:= if { other list of conditions }, re-enable <to-be-recovered> when the to-be-recovered alert is triggered, it will send any notifications that are configured for it. it might be useful to periodically push those notifications out again and again until the alert definition is recovered. we can extend recovery-based alert definition to have a new option: re-notify every X units of Y time, until alert is recovered is re-enabled then, when the to-be-recovered alert is triggered, we can create a periodic quartz job that will fire every X units of Y time. when it fires, it will look up the current notification attached to the alert definition that created it, and re-notify those persons / devices. when the recovery alert eventually triggers, it will unschedule this quartz job to stop the notifications.
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-1207