Monitor Azure Stack Hub hardware components
The Azure Stack Hub health and monitoring system monitors the status of the storage subsystem and raises alerts as needed. The health and monitoring system can also raise alerts for the following hardware components:
- System fans
- System temperature
- Power supply
- CPUs
- Memory
- Boot drives
Note
Before you enable this feature, you must validate with your hardware partner that they're ready. Your hardware partner will also provide the detailed steps for enabling this feature in the baseboard management controller (BMC). The user encryption in the base board management controller must be set to AES for build 2005 and later.
SNMP listener scenario
An SNMP v3 listener is running on all three ERCS instances on TCP port 162. The BMC must be configured to send SNMP traps to the Azure Stack Hub listener. You can get the three PEP IPs from the administrator portal by opening the region properties view.
Sending traps to the listener requires authentication and must use the same credential as accessing base BMC itself.
When an SNMP trap is received on any of the three ERCS instances on TCP port 162, the OID is matched internally and an alert is raised. The Azure Stack Hub health and monitoring system only accepts OIDs defined by the hardware partner. If an OID is unknown to Azure Stack Hub, it won't match it to an alert.
Once a faulty component is replaced, an event is sent from the BMC to the SNMP listener that indicates the state change. The alert then closes automatically in Azure Stack Hub.
Note
Existing alerts do not close automatically when the entire node or motherboard is replaced. The same applies when the BMC loses its configuration; for example, due to a factory reset.