2 SNMP agent

Overview

You may want to use SNMP monitoring on devices such as printers, network switches, routers or UPS that usually are SNMP-enabled and on which it would be impractical to attempt setting up complete operating systems and Zabbix agents.

To be able to retrieve data provided by SNMP agents on these devices, Zabbix server must be initially configured with SNMP support by specifying the --with-net-snmp flag.

SNMP checks are performed over the UDP protocol only.

Zabbix server and proxy daemons query SNMP devices for multiple values in a single request. This affects all kinds of SNMP items (regular SNMP items, SNMP items with dynamic indexes, and SNMP low-level discovery) and should make SNMP processing much more efficient. See the bulk processing section for technical details on how it works internally. Bulk requests can also be disabled for devices that cannot handle them properly using the "Use bulk requests" setting for each interface.

Zabbix server and proxy daemons log lines similar to the following if they receive an incorrect SNMP response:

SNMP response from host "gateway" does not contain all of the requested variable bindings

While they do not cover all the problematic cases, they are useful for identifying individual SNMP devices for which bulk requests should be disabled.

Zabbix server/proxy will always retry at least one time after an unsuccessful query attempt: either through the SNMP library's retrying mechanism or through the internal bulk processing mechanism.

Configuring SNMP monitoring

To start monitoring a device through SNMP, the following steps have to be performed:

Step 1

Find out the SNMP string (or OID) of the item you want to monitor.

To get a list of SNMP strings, use the snmpwalk command (part of net-snmp software which you should have installed as part of the Zabbix installation) or equivalent tool:

snmpwalk -v 2c -c public <host IP> .

As '2c' here stands for SNMP version, you may also substitute it with '1', to indicate SNMP Version 1 on the device.

This should give you a list of SNMP strings and their last value. If it doesn't then it is possible that the SNMP 'community' is different from the standard 'public' in which case you will need to find out what it is.

You can then go through the list until you find the string you want to monitor, e.g. if you wanted to monitor the bytes coming in to your switch on port 3 you would use the IF-MIB::ifHCInOctets.3 string from this line:

IF-MIB::ifHCInOctets.3 = Counter64: 3409739121

You may now use the snmpget command to find out the numeric OID for 'IF-MIB::ifHCInOctets.3':

snmpget -v 2c -c public -On <host IP> IF-MIB::ifHCInOctets.3

Note that the last number in the string is the port number you are looking to monitor. See also: Dynamic indexes.

This should give you something like the following:

.1.3.6.1.2.1.31.1.1.1.6.3 = Counter64: 3472126941

Again, the last number in the OID is the port number.

In the last example above value type is "Counter64", which internally corresponds to ASN_COUNTER64 type. The full list of supported types is ASN_COUNTER, ASN_COUNTER64, ASN_UINTEGER, ASN_UNSIGNED64, ASN_INTEGER, ASN_INTEGER64, ASN_FLOAT, ASN_DOUBLE, ASN_TIMETICKS, ASN_GAUGE, ASN_IPADDRESS, ASN_OCTET_STR and ASN_OBJECT_ID. These types roughly correspond to "Counter32", "Counter64", "UInteger32", "INTEGER", "Float", "Double", "Timeticks", "Gauge32", "IpAddress", "OCTET STRING", "OBJECT IDENTIFIER" in snmpget output, but might also be shown as "STRING", "Hex-STRING", "OID" and other, depending on the presence of a display hint.

Step 2

Create a host corresponding to a device.

Add an SNMP interface for the host:

• Enter the IP address/DNS name and port number
• Select the SNMP version from the dropdown
• Add interface credentials depending on the selected SNMP version:
  • SNMPv1, v2 require only the community (usually 'public')
  • SNMPv3 requires more specific options (see below)
• Leave the Use bulk requests checkbox marked to allow bulk processing of SNMP requests

In case of wrong SNMPv3 credentials (security name, authentication protocol/passphrase, privacy protocol):

• Zabbix receives an ERROR from net-snmp, except for wrong Privacy passphrase in which case Zabbix receives a TIMEOUT error from net-snmp;
• (since Zabbix 6.0.13) SNMP interface availability will switch to red (unavailable).

You can use one of the provided SNMP templates (Template SNMP Device and others) that will automatically add a set of items. However, the template may not be compatible with the host. Click on Add to save the host.

Step 3

Create an item for monitoring.

So, now go back to Zabbix and click on Items for the SNMP host you created earlier. Depending on whether you used a template or not when creating your host, you will have either a list of SNMP items associated with your host or just an empty list. We will work on the assumption that you are going to create the item yourself using the information you have just gathered using snmpwalk and snmpget, so click on Create item. In the new item form:

• Enter the item name
• Change the 'Type' field to 'SNMP agent'
• Enter the 'Key' as something meaningful
• Make sure the 'Host interface' field has your switch/router in it
• Enter the textual or numeric OID that you retrieved earlier into the 'SNMP OID' field, for example: .1.3.6.1.2.1.31.1.1.1.6.3
• Set the 'Type of information' to Numeric (unsigned)
• Enter an 'Update interval' and 'History storage' period if you want them to be different from the default
• In the Preprocessing tab, add a Change per second step (important, otherwise you will get cumulative values from the SNMP device instead of the latest change). Choose a custom multiplier if you want one.

All mandatory input fields are marked with a red asterisk.

Now save the item and go to Monitoring → Latest data for your SNMP data!

Example 1

General example:

Note that OID can be given in either numeric or string form. However, in some cases, string OID must be converted to numeric representation. Utility snmpget may be used for this purpose:

snmpget -On localhost public enterprises.ucdavis.memory.memTotalSwap.0

Example 2

Monitoring of uptime:

Internal workings of bulk processing

Zabbix server and proxy query SNMP devices for multiple values in a single request. This affects several types of SNMP items:

• regular SNMP items
• SNMP items with dynamic indexes
• SNMP low-level discovery rules

All SNMP items on a single interface with identical parameters are scheduled to be queried at the same time. The first two types of items are taken by pollers in batches of at most 128 items, whereas low-level discovery rules are processed individually, as before.

On the lower level, there are two kinds of operations performed for querying values: getting multiple specified objects and walking an OID tree.

For "getting", a GetRequest-PDU is used with at most 128 variable bindings. For "walking", a GetNextRequest-PDU is used for SNMPv1 and GetBulkRequest with "max-repetitions" field of at most 128 is used for SNMPv2 and SNMPv3.

Thus, the benefits of bulk processing for each SNMP item type are outlined below:

• regular SNMP items benefit from "getting" improvements;
• SNMP items with dynamic indexes benefit from both "getting" and "walking" improvements: "getting" is used for index verification and "walking" for building the cache;
• SNMP low-level discovery rules benefit from "walking" improvements.

However, there is a technical issue that not all devices are capable of returning 128 values per request. Some always return a proper response, but others either respond with a "tooBig(1)" error or do not respond at all once the potential response is over a certain limit.

In order to find an optimal number of objects to query for a given device, Zabbix uses the following strategy. It starts cautiously with querying 1 value in a request. If that is successful, it queries 2 values in a request. If that is successful again, it queries 3 values in a request and continues similarly by multiplying the number of queried objects by 1.5, resulting in the following sequence of request sizes: 1, 2, 3, 4, 6, 9, 13, 19, 28, 42, 63, 94, 128.

However, once a device refuses to give a proper response (for example, for 42 variables), Zabbix does two things.

First, for the current item batch it halves the number of objects in a single request and queries 21 variables. If the device is alive, then the query should work in the vast majority of cases, because 28 variables were known to work and 21 is significantly less than that. However, if that still fails, then Zabbix falls back to querying values one by one. If it still fails at this point, then the device is definitely not responding and request size is not an issue.

The second thing Zabbix does for subsequent item batches is it starts with the last successful number of variables (28 in our example) and continues incrementing request sizes by 1 until the limit is hit. For example, assuming the largest response size is 32 variables, the subsequent requests will be of sizes 29, 30, 31, 32, and 33. The last request will fail and Zabbix will never issue a request of size 33 again. From that point on, Zabbix will query at most 32 variables for this device.

If large queries fail with this number of variables, it can mean one of two things. The exact criteria that a device uses for limiting response size cannot be known, but we try to approximate that using the number of variables. So the first possibility is that this number of variables is around the device's actual response size limit in the general case: sometimes response is less than the limit, sometimes it is greater than that. The second possibility is that a UDP packet in either direction simply got lost. For these reasons, if Zabbix gets a failed query, it reduces the maximum number of variables to try to get deeper into the device's comfortable range, but (starting from 2.2.8) only up to two times.

In the example above, if a query with 32 variables happens to fail, Zabbix will reduce the count to 31. If that happens to fail, too, Zabbix will reduce the count to 30. However, Zabbix will not reduce the count below 30, because it will assume that further failures are due to UDP packets getting lost, rather than the device's limit.

If, however, a device cannot handle bulk requests properly for other reasons and the heuristic described above does not work, since Zabbix 2.4 there is a "Use bulk requests" setting for each interface that allows to disable bulk requests for that device.