[EDGENT-14] Notify external systems on recovery from failure

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Runtime
• Labels: failure-recovery
• Assignee: Unassigned
• Reporter: Victor Dogaru
• Created: 08/Mar/16 21:36
• Updated: 08/Mar/16 21:36

Description

Notify external systems (for example an IoT message hub) when a Quarks application or system recovers from failure. This may include one or more of the following mechanisms:

• Quarks system application which polls a monitor service and pushes Job state to remote IoT management systems. The monitor service (described in QUARKS-8) restarts topologies in case of crash.
• Provide health attribute to JobMXBean and support attribute change JMX notifications (requires QUARKS-12).
• System watchdog which restarts a Quarks service in case of crash (ties with QUARKS-13) notifies the device owner/administrator. This would require a notification mechanism external to Quarks (such as email?).