Warnings and Notifications#

On console sign-in, a health report indicates the system status. This health report shows this data:

Last login: Tue Sep 3 10:19:07 2013 from 172.29.232.68
host: alan, role: standalone, load: 0.35, USERS: 3
date: 2013-09-03 10:20:02 +00:00, up: 2:05
network: 172.29.89.182, ntp: 172.29.1.15
SECURITY UPDATES: 136 security updates available
database: 8.0Gb
services: ok

The report values mean:

  • last console sign-in and IP address source

  • the load average of the system

  • the number of users currently signed in

  • the system uptime

  • the status of the system services

  • whether security updates are available

  • disk, CPU, and memory warnings if applicable

  • warnings are displayed in uppercase to draw attention

The report can be redisplayed by typing the command:

health

The system can be configured to forward warnings and notifications to various destinations, including:

  • local email

  • remote email addresses

  • remote SNMP destinations

Local email allows the administrator to view a list of warnings, and delete them as necessary.

Note

If this configuration is carried out, it should be set on each node.

Typically, each node will be configured with its own sender email address to identify the source node (for example unified_node_1@customer.domain).

Additionally,

  • nodes can be configured with different emailrelay servers, for example, according to data center

  • nodes can be configured with different SNMP servers and error levels

The notification destinations can be displayed with notify list. The destinations for each event level can be set with notify add info|warn|error <destination-URI> Refer to the Network URI Specification topic for a detailed description of URIs.

Note that email notifications require the mail relay to be set with notify emailrelay <smtp ip address>[:port]. Add the SMTP IP address and optionally an allowed port (default port is 25).

The default from address is platform@<hostname>. A from outgoing address for the email relay can also be configured with notify emailfrom <from: email address> - see: Enable Health Monitoring. A test event can be generated with notify test info|warn|error to test the notification delivery mechanism.

Examples:

  • notify add info mailto:sysadmin@mycompany.com

  • notify add error snmp://public@mysnmpserver.com

$ notify add error snmp://[email protected]
   notifications:
     emailrelay: 172.1.1.1
     level:
         error:
             snmp://[email protected]
             mailto:platform@localhost
         info:
             mailto:platform@localhost
         warn:
             mailto:platform@localhost

In addition to external email and SNMP alerts, the system also records various events to a local mailbox. Refer to the Mail Command section for details.

SNMP CPU load notifications are set using:

snmp load <1min load> <5min load> <15min load>

This results in notifications being sent should the threshold be exceeded. For a server with 2 CPUs, it is recommended that this setting be:

snmp load 8 4 2

This means that notifications are sent if the 2-CPU system load averages over the last 1, 5, and 15 minutes reach these values.