When the managed system generates certain events, it will forward a SNMP trap. The reason for the event trap is contained in the SNMP MIB string. Note that if the corresponding SNMP MIB is not loaded on the NMS, a numerical representation of the SNMP entry is provided. The list of monitored events is described in the SNMP Trap section below. A detailed breakdown of each SNMP trap type is provided in the appendix.
The SNMP will send traps to the trap destination configured. If the trap destination is incorrect or not configured, the NMS will not receive the traps.
The following system parameters are monitored by default:Disk Space: warnings are issued if the file system becomes full
System Load Monitoring: warnings are issued if the system load is excessive (the system load parameters can be defined during configuration)
SNMP: standard SNMP System Events, for example, Cold Start
Process state changes: Informative messages are sent to the NMS indicating that processes have been restarted.
In general, the originator of the SNMP traps is determined by originating hostname / IP address. Many Network Management Systems provide trap management and escalation per system being managed, including identification based on system name, location and contact details. Those events monitored directly by Cisco Unified
Communications Domain Manager 10.6(1) (e.g. disk space, system load and process warnings) include the system name as part of the variable bindings to assist identification of the originating system.
Disk Space Low [High Priority]Priority: HIGH
Action: Call support
SNMP monitors the percentage of free space available, and will raise a trap if the filesystem becomes full. By default, the threshold is 10% free space available. The filesystem is used to store log files, etc. and under normal conditions will recycle these to ensure that disk space is managed. If a disk space low warning is received, it should be treated as a high priority, customer support should be contacted to determine the reason for the filesystem becoming full.
Excessive Load [Medium Priority]
During configuration, the maximum load can be specified as an average over 1minute, 5minute and 15minute intervals. The system load may spike during certain activities such as bulk loading and will recover. The warning should only be treated as serious if the system load is high for an extended period (e.g. over a 10 minute average).
The load can be monitored either on the NMS for further excessive load traps, or via the command line interface (documented in the Command line interface guide) using the status and healthlog commands.
If the 15-minute threshold is exceeded, diagnosis is required to determine the cause of the high load. Ensure that sufficient CPU and Memory resources are available to the Cisco Unified
Communications Domain Manager 10.6(1) system as per the initial hardware scaling requirements. The problem may also be caused by incidents such as network outages, delayed backups, manual intervention by a system administration, etc. Please contact support for further diagnosis of the problem.
Process State Changes [Medium Priority]Priority: Medium
Action: Monitor
Process state changes (e.g., restart of services) are normal behavior when the Cisco Unified
Communications Domain Manager 10.6(1) system is started or shutdown. Manual intervention by a system administrator is also likely to cause services to start or stop. The system manages processes automatically and will restart services as required. Process state changes are normal during HA (cluster failover).
The state of processes can be monitored either on the NMS for subsequent process state change traps, or via the command line interface (documented in the Command line interface guide) using the monitor command.
Check if there is a known outage, change control window, or scheduled work in progress for this platform. If there is none, then these traps represent a high priority issue and need to be logged with support.
Standard SNMP Events [Low Priority]Priority: Low
Action: Monitor
Standard SNMP traps for cold-start and shutdown are generated by the Cisco Unified
Communications Domain Manager 10.6(1) system when it is started or shutdown. Manual intervention by a system administrator may also generate these traps if the system is restarted. Cold-start notices may also indicate HA (cluster failover).
The state of the Cisco Unified
Communications Domain Manager 10.6(1) system can be monitored either on the NMS for subsequent cold-start traps, or via the command line interface (CLI) using the monitor and healthlog commands.
Check if there is a known outage, change control window, or scheduled work in progress for this platform. If there is none, then these traps represent a high priority issue and need to be logged with support.
Reconfigure SNMP
SNMP configuration settings can be managed from the CLI. Refer to the CLI notify command:platform@development:~$ notify
USAGE:
------
notify add [info|warn|error] - Add the email or snmp URI to a
<email/snmp-uri> ... specified notification level.
SNMP must be configured under the SNMP menu and the SNMP URI needs to be configured for all the notify severity levels(info|warn|error]).
The following options can be configured under the SNMP menu in the CLI.Enabled -Enable or disable SNMP Queries
Community- SNMP v2c Community String used to query this server
Authorized Username - SNMP v3 Username to query this server
Password - SNMP v3 Password to query this server
Query - IP address that is allowed to query this server
Sysname - Name of this server, as it will appear when queried via SNMP
Syslocation - Location of this server
Syscontact - Contact person(s) for this server (email address)
Load1 - 1 Minute load average alarm value
Load5 - 5 Minute load average alarm value
Load15 - 15 Minute load average alarm value
Two SNMP Trap destinations can be configured:
The following options can be configured in the CLI:Hostname - Server name to send SNMP traps to.
Version - Version of SNMP to use for sending trap, version 2c or 3.
Community - refer to the SNMP-URI command usage.
Mode - Send Trap or Inform message.
Username - refer to the SNMP-URI command usage.
Password - refer to the SNMP-URI command usage.
Encryption - refer to the SNMP-URI command usage.
Engineid - To send traps as. - Currently not implemented