AN41 - Power monitoring over SNMP in Zabbix / Nagios for NETIO power sockets

Supported devices: NETIO 4All, PowerPDU 4C, PowerDIN 4Pz, PowerBOX 4Kx, PowerPDU 8QS

The AN41 Application Note demonstrates the use of a SNMP monitoring system and shows how to monitor power in IT and other applications.

Problem statement

For our AN41 demonstration, let us take a NETIO PowerPDU 4C with the following appliances connected to its outputs:

• Output1 - Fridge
• Output2 - Kettle
• Output3 - Coffee machine
• Output4 - Microwave oven

Goals for AN41

• Measure the consumption at all sockets and display it in graphs.
• Send an e-mail notification whenever the kettle is turned on.
• Send an e-mail warning whenever the fridge door is left open.
• Send an e-mail warning if the fridge stops working.

The PowerPDU 4C is connected to a LAN

• IP address:  192.168.101.31
• SNMP: Read permissions for the SNMP community:  netioVkuchynce7523
• Our PC and the monitoring server with Zabbix can access the NETIO device over SNMP

How SNMP works

SNMP is a protocol for remote monitoring and control of devices. With NETIO smart power sockets, SNMP can be used to read OID object values (socket state (0/1), the immediate current [mA], and the consumption counter [Wh]).

By writing values over SNMP, the output can be controlled (switched on/off, toggled, briefly switched on or off). However, we recommend to use the more secure SNMP v3 for writing values.

Values are identified by their OID in a hierarchical tree.

A variable can be identified with its number or its textual name.

In AN41, we will be using the numeric OIDs. Although text identifiers are clearer, they require importing the NETIO MIB into the monitoring software. Not all SW tools support MIBs.

In the hierarchical tree of OID values, NETIO has the following assigned prefix/path: .1.3.6.1.4.1.47952.

For a detailed description of SNMP in NETIO products, including examples and guides for installing the utilities, see AN11: AN11 SNMP management of 110/230V power outlets from the command line in Windows and Linux

What to measure: Wh or A / W ?

Two kinds of values can be used for energy monitoring:

1) Consumption counter [Wh] - measures the total energy transferred through the socket (netioOutputEnergy [Wh])

• Accurate determination of the total consumption in Wh
• NETIO device takes care about calculating the consumed energy
• Less demanding on the monitoring system (it is possible to measure just once a day and the measurement will still be accurate)
• Less clear and understandable values in graphs: What is shown is either the total consumption that is constantly rising; or more often, the increase in total consumption during a unit of time, such as Wh/s.
• Measurement accuracy depends on the internal measurement accuracy of the NETIO device, it is not distorted by the monitoring and reading.
• The reading is in Wh, which limits the ability to measure and evaluate short-term loads or small loads. The value is an integer in Wh: for example, with a 50 W load, the value increases by 1 every 72 seconds.

2) Immediate current [A] - (netioOutputCurrent [mA] or netioOutputLoad [W]), the total energy consumption is calculated by the monitoring software

• To monitor circuit breaker capacity, the immediate value is more suitable (netioOutputCurrent [mA] or netioOutputLoad [W])
• Some graphs of current and some results may be easier for the users to understand
• When measuring, the values need to be read at least twice as often as is their rate of change. Most monitoring software reads once in 5 minutes, which is only sufficient in some cases.
• If the load is active only for a short time, or the consumption often fluctuates, the value needs to be read more frequently than it changes. For example, if the period of high consumption lasts 30s (e.g. the time to bring water in the kettle to boil), and we read the values once every 5 minutes (default for most monitoring systems), most likely the consumption will not be recorded at all.

In most situations, we recommend to combine the measurement of both kinds of values.

The general measurement accuracy of NETIO smart socket devices is better than 1% in case of Power PDU and PowerCable REST 101x, and better than 2% in case of Netio 4.

Note: If writing is also needed (sockets need to be controlled), SNMP v3 must be used.

Reading over SNMP using the command line and Net-SNMP utilities

This method of reading values is useful for checking that things work, or for simple usage in scripts.

The following examples use Net-SNMP on Linux. Examples for Windows are shown in AN11: AN11 SNMP management of 110/230V power outlets from the command line in Windows and Linux

Syntax (SNMP v1):

  snmpget -v 1 -c <SNMP-community> <IP> <OID>

$ snmpget -v 1 -c netioVkuchynce7523 192.168.101.31 .1.3.6.1.4.1.47952.1.1.1.28.1

iso.3.6.1.4.1.47952.1.1.1.28.1 = INTEGER: 17

The value of 17 means that the immediate current drawn from socket no. 1 is 17 mA.

snmpget -v 1 -c netioVkuchynce7523 192.168.101.31 .1.3.6.1.4.1.47952.1.1.1.26.1

iso.3.6.1.4.1.47952.1.1.1.26.3 = INTEGER: 25132

The value of 25132 means that the total energy consumed through socket no. 3 is 25132 Wh (the start time of the measurement can be read from OID 27.3).

If SNMP v3 is used:

$ snmpget -v 3 -a SHA -A onlinedemo -l authPriv -u onlinedemo -x AES -X onlinedemo netio-4c.netio-products.com iso.3.6.1.4.1.47952.1.1.1.28.1

iso.3.6.1.4.1.47952.1.1.1.28.1 = INTEGER: 2577

Zabbix SNMP monitoring

The Zabbix documentation includes a detailed description of how to monitor custom OIDs:

https://www.zabbix.com/documentation/current/manual/config/items/itemtypes/snmp

Configuration -> Hosts -> Create host

• Host tab
  • "Host name": netio-kuchynka
  • "Groups":    any Zabbix group
  • "Agent Interfaces": Remove
  • "SNMP interfaces": IP: 192.168.101.31

• Templates tab (for optional ICMP ping monitoring that the Netio device is alive):
  • Link new Templates: "Template Module ICMP Ping"
    • Add
    • Update

Configuration -> Hosts -> netio-kuchynka, Graphs tab

• "Create graph"
  •     Name:    netio-kuchynka-OutputEnergy
  •     Height: 300
  •     Y axis MIN value: Fixed, 0
  • Items: Add:
    • netio-kuchynka-SNMP-Output1-OutputEnergy
    • netio-kuchynka-SNMP-Output2-OutputEnergy
    • netio-kuchynka-SNMP-Output3-OutputEnergy
    • netio-kuchynka-SNMP-Output4-OutputEnergy

• We can immediately see what the graph looks like in "Preview".
• Add

Repeat the graph creation steps for Current and Load.

Monitoring -> Screens

• Create screen
  • Name: netio-kuchynka-SNMP
  • Columns: 1
  • Rows: 3
• Add
• In the list of created Screens with the „Name“ of netio-kuchynka-SNMP, click „Actions“: Constructor
  • Click „Change“
    • Graph: Click „Select“ to select the desired graph
      • Group: Zabbix group where the “netio-kuchynka” host was created
      • Host: netio-kuchynka
    • netio-kuchynka-SNMP-Current
    • Add
  • SImilarly, add the 2 other graphs for:
    • netio-kuchynka-SNMP-Load
    • netio-kuchynka-SNMP-OutputEnergy

The Screen with graphs can be displayed in Monitoring-> Sreens -> “netio-kuchynka-SNMP”.

The Current graphs are in mA, the Load graphs are in Watts. The polling period is 30s (Zabbix default). The 2 kW consumption of the kettle at Output2 is clearly visible. When the time scale is changed, the values are recalculated and the averages for the given time are displayed. For example, when displaying the history for the last 3 months, the average kettle consumption will only amount to about 100 W.

The OutputEnergy graph is in Wh/s, corresponding to the measurement method – what is measured is the total consumption since the start of the measurement (typically corresponds to when the NETIO device was restarted). The advantage is that the sampling period is the fastest supported by the device (2 seconds for NETIO 4x) and that the measurement continues even if the monitoring server or NETIO connectivity is down.

Nagios SNMP monitoring

The most common method of using SNMP is with the check_snmp plugin (script).

From the command line, measurement result can be obtained like this (example for Output1 Load):

/usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.1

SNMP OK - 0 | iso.3.6.1.4.1.47952.1.1.1.25.2=0

The script returns that the OID value is 0 (zero consumption at outlet 2 at the time of measurement). From a script/plugin, Nagios expects a return value indicating that the status is OK, Warning or Critical. Usually, this involves comparing the OID value with a benchmark. The comparison is done by the check_snmp script that takes the following arguments:

            -w <Warning range>

            -c <Critical range>

The ranges indicate when the alert is NOT generated. Attention – the logic of using ranges in check_snmp plugin calls reversed to that of Nagios and to what one would expect (for details, see https://wiki.xdroop.com/space/nagios/check_snmp).

Note: In Nagios, the script name is always check_snmp; however, the path may be different in different distributions or systems. The /usr/lib/nagios/plugins/check_snmp path is valid for Ubuntu 18, it may be different on your operating system – check the installation path for files from the monitoring-plugins-standard package.

Note: We do NOT recommend using the check_netio Nagios plugin for NETIO 4 and newer devices. The plugin uses kshell access and it was developed and tested only for the older NETIO 230 devices.

For the above problem statement, Nagios is not the most suitable because the default installation is relatively limited and does not support graphs.

The advantage of Nagios is its simplicity and extensibility with scripts. In the default installation, it is easy to monitor if individual counter values are exceeded. Therefore, let us simplify the problem statement for Nagios:

• The consumption at Output2 is greater than 2000 W. Nagios measures a value only once every 5 minutes (so that we need to “catch” the moment when the kettle is on); however, we will ignore this problem in this example and boil water in the kettle 3 times in a row.
• Detect if the coffee machine is unplugged or out of order. The coffee machine draws an idle current of 15mA. It is not useful to measure the power input [W] because it is rounded to 0 W. So, we will measure the consumption in mA.

Command-line measurement test:

• When the kettle is off:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.2

SNMP OK - 0 | iso.3.6.1.4.1.47952.1.1.1.25.2=0

• When the kettle is off:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.2

SNMP OK - 2202 | iso.3.6.1.4.1.47952.1.1.1.25.2=2202

• The coffee machine is plugged in and idle:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3

SNMP OK - 15 | iso.3.6.1.4.1.47952.1.1.1.28.3=15

• The coffee machine is making coffee:

/usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3

SNMP OK - 5259 | iso.3.6.1.4.1.47952.1.1.1.28.3=5259

• The coffee machine is unplugged or out of order:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3

SNMP OK - 0 | iso.3.6.1.4.1.47952.1.1.1.28.3=0

We set the detection ranges and test them:

• The kettle is on and generates a critical event

/usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.2 -w :10 -c :2000

SNMP CRITICAL - *2165* | iso.3.6.1.4.1.47952.1.1.1.25.2=2165;:10;:2000

• The kettle is off and the status is OK:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.2 -w :10 -c :2000

SNMP OK - 0 | iso.3.6.1.4.1.47952.1.1.1.25.2=0;:10;:2000

• The coffee machine draws more than 10mA and the status is OK:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3 -w 10: -c 5:

SNMP OK - 14 | iso.3.6.1.4.1.47952.1.1.1.28.3=14;10:;5:

• The coffee machine is unplugged and the status is critical:

root@sandbox:~# /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3 -w 10: -c 5:

SNMP CRITICAL - *0* | iso.3.6.1.4.1.47952.1.1.1.28.3=0;10:;5:

We will save the configuration corresponding to the above to the following file

/etc/nagios3/conf.d/netio.cfg

File contents:

define command{

        command_name    check_snmp

        command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C netioVkuchynce7523 -o $ARG1$ -w $ARG2$ -c $ARG3$

}

define host{

        use                     generic-host

        host_name               netio

        alias                   netio

        address                 192.168.101.31

}

define service{

        use                             generic-service

        host_name                       netio

        service_description             NetioOutput1Load

        check_command                   check_snmp!.1.3.6.1.4.1.47952.1.1.1.25.2!:10!:2000

        # /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.25.2 -w :10 -c :2000

        # warning above 10W

        # critical above 2000W

}

define service{

        use                             generic-service

        host_name                       netio

        service_description             NetioOutput3Current

        check_command                   check_snmp!.1.3.6.1.4.1.47952.1.1.1.28.3!10:!5:

        # /usr/lib/nagios/plugins/check_snmp -H 192.168.101.31 -C netioVkuchynce7523 -o .1.3.6.1.4.1.47952.1.1.1.28.3 -w 10: -c 5:

        # warning below 10mA

        # critical below 5mA

}

Restart Nagios:

service nagios3 restart

There should be two more services in the web interface that monitor the ranges defined above:

Note: More information about Nagios

Supported FW for NETIO 4x:

3.1.0 and higher (Archive firmware)