An example of this system in action:
An ATM reports that it is "Out of Service" over RMS from a dispenser error. IMS then relays this information and the parsed error through SMS to a user. This user is then provided an option to Reboot the ATM. The user sends the reboot command and IMS tries RMS and Hard Reboots. The ATM restarts and goes back into service, IMS detects this and a message is sent to the user stating that the issue is resolved. When Auto-Healing is enabled, this entire process is handled by Hercules AI and the user is just informed when the incident has been resolved.
This flow can be applied to other scenarios like theft alarms with the ability to disarm and trigger sirens or custom low cash levels with the ability to assign users who are in charge of the fill and how much to add in. This system is continually learning and eventually IMS will have logic to detect issues before they happen.
IMS relies heavily on schedules to ensure that notifications are being sent to the right people at the right times. Schedules can allow you to fine tune the system to your needs or can be left on default for ease of use. You can edit schedules in the Schedules Panel.
Rules:
- Incidents are triggered from alarms from your devices sensors, RMS data, Insights data, and Network Activity.
- Repeated alarms from the same device are grouped into one incident until they are resolved.
- Notifications are sent when incidents are validated. Incidents will monitor and delay sending notifications until they are validated. These notifications allow you to respond to the incident by sending commands to Reboot, Disarm, Acknowledge, or Mark as Resolved.
- Notifications for incidents can be managed in the Schedules Panel. Schedules define who receives incident notifications based on device groups and alarms.
- When an incident is acknowledged or resolved, all users who are part of the incident will be notified.
- Acknowledging an incident assigns it to the user who acknowledged it. This can be changed in the detailed view of each incident.
- Incidents can auto-resolve if the condition clears. For instance, an "Out of Cash" incident will auto-resolve once it reports a new bill count.
- Incidents will be marked as "Timed Out" and a notification will be sent after 24 hours if they have not been cleared. Alarms defined as "Urgency - High" will only timeout after 7 days.
Noise Filter:
The Noise Filter removes any incidents from the view that Hercules AI has determined are benign or have automatically resolved. These incidents can be temporary errors or inactivity reported to RMS, Processors, or the device itself. Turning off the Noise Filter will show incidents that are filtered or are currently being monitored.
Hercules AI:
Incidents labeled as with the Hercules AI badge have been monitored, evaluated, and acted upon by Hercules AI. This includes detecting inactivity patterns, temporary errors, cash loads, RMS and transaction errors. Hercules AI learns to determine what are actionable issues and notifies you accordingly, cutting out the rest to reduce false alarms. You can view most of these by disabling noise filters.
Comments
0 comments
Please sign in to leave a comment.