Using CA AppLogic › Grid User Guide › Monitoring Grids › Monitor using SNMP

Monitor using SNMP

Simple Network Management Protocol (SNMP) is a protocol for monitoring devices over IP networks.

The grid provides standard statistics on Linux hosts for each physical server and the grid controller. In addition, you can access CPU, memory, disk, and network usage statistics for all running appliances and the grid controller, as reported by the hypervisor. Network and disk names that display their AppLogic names, such as terminal 'in' or volume 'boot'. Appliance statistics are provided as an SNMP extension.

No additional grid resources, such as dedicated monitoring appliances or applications, are required. There are no changes needed to your applications and nothing to enable per application or appliance.

The SNMP data is accessible from the external network of the grid controller over a secure, authenticated connection. AppLogic supports only SNMPv3.

Enable SNMP Monitoring

You can enable SNMP monitoring when building a new grid as well as on an existing grid. By default, SNMP monitoring is disabled.

To enable SNMP monitoring:

• New Grid - refer to the Grid Parameters section of the BFC User Guide.
• Existing Grid - use the following CLI command provided in the grid set:

  grid set snmp_enabled=<0|1> snmp_user=<user> snmp_password=<password>
  

You must specify both the username and password. The password must be minimum of 8 characters and a maximum of 24 characters.

For information on verifying the current SNMP state, such as enabled or disabled, in the output of 'grid info', refer to the Grid Set section of the Command Line Reference guide.

Get Statistics

You can monitor the entire grid from a single point on the grid controller over a secure connection on its external network.

Command Example

# Get stats for the grid controller host

snmpwalk -v3 -u "MyUser" -a MD5 -A "MyPassword" -l authPriv  -x DES -X "MyPassword" controller_name_or_ip

# Get stats for a server (srv1), will include stats for all appliances on that server as well as standard snmp stats

snmpwalk -n "srv1" -v3  -u "MyUser" -a MD5 -A "MyPassword" -l authPriv  -x DES -X "MyPassword" controller_name_or_ip

# Get stats for all appliances running on a server (srv1)

snmpwalk -n "srv1" -v3 -a MD5 -A $pwd  -l authPriv  -x DES -X "MyPassword" -u "MyUser" controller_name_or_ip  NET-SNMP-EXTEND-MIB::nsExtendOutputFull

# Get stats for all servers, OIDs 1.3 from each server will be mapped to 1.4. on the controller

snmpwalk -v3 -u "MyUser" -a MD5 -A "MyPassword" -l authPriv  -x DES -X "MyPassword" controller_name_or_ip 1.4

Response Example

Appliance statistics are returned as an UDL formatted string. Memory, network, and disk values are in bytes. CPU values are in seconds.

Note: The network statistics for shared interfaces report the same data for all terminals that are sharing the same interface.

For example:

vm Domain-0
   {
   memory = 1073336320
   cpu = 625115.53
   }
vm controller
   {
   memory = 1073741824
   cpu = 396621.07
   iface : name = "internal", rx = 3499828357, tx = 308441339
   iface : name = "external", rx = 76663726, tx = 95916570
   volume : name = "boot", read = 263594, write = 19315084
   volume : name = "meta", read = 7621, write = 1956036
   volume : name = "impex", read = 11563453, write = 13636232
   volume : name = "swap", read = 1568, write = 1132
   }
vm vm.srv1.pg.main.lux6
   {
   memory = 134217728
   cpu = 1773.11
   iface : name = "in", rx = 0, tx = 670
   iface : name = "default", rx = 498883743, tx = 408949762
   iface : name = "mon", rx = 0, tx = 670
   iface : name = "net", rx = 0, tx = 670
   iface : name = "out", rx = 0, tx = 670
   volume : name = "boot", read = 18408, write = 22841
   volume : name = "usr", read = 9123, write = 0
   }