v0.04 (in progress - beta 1)
- enhancement: multiple interfaces with same descr/physical address handling
- enhancement: unified the GetTablexxx snmp functions
- enhancement: code simplification
- enhancement: revamped generated html structure, introduced buttons, all styles in css
- enhancement: better node reboot handling
- enhancement: new plugin version detection
- enhancement: new state file format
- enhancement: logic now based on a generated ifName instead of index
- enhancement: improved vlan property checks
- enhancement: code restructuration and cleaning
- enhancement: interface duplex mode collect now more robust
- enhancement: 32/64bits counter change detection,
- enhancement: 5 debug levels
- feature: added a snmp oid recorder for device simulation purpose (for testing phase). win32 precompiled for easy use, and howto
- feature: netmask conversion from quad to dec notation
- feature: added spanning tree state tracking (cistp only)
- feature: much better performances by retrieving tables/bulks instead of single values
- feature: batch files to simplify taking device snapshot for simulation
- feature: nodetype option handling cisco, hp, and netscreen node types
- feature: Juniper Netscreen support (zone,vsys,mng,IPs)
- feature: support of snmp transport domain udp, tcp, udp6 and tcp6
- bugfix: fixed space problem in action column on ie/opera
- bugfix: undef speed fix
v0.03-2 (not published)
- bugfix: workaround for devices with same interface name 'Ethernet Interface'
v0.03-1
- enhancement: interface coming from 'empty' operstatus to another operstatus does not trigger an alert of operstatus change (ex: newly added interface)
- feature: introducing ifdescr preformating, applied to Cisco ASA specific interfaces
- bugfix: remove single quote from perfdata names and url to pnp
- bugfix: fixed major bug - config option -wp and -cp now honored (Birger Schmidt)
- enhancement: negative value for err/disc pkt tracking disable the specific tracking.
ex: --warning 50,0,-1 --critical 80,100,-1 disable discard packet tracking
- enhancement: better support of graphing solutions
- add: new nagiosgrapher templates
v0.03
- feature: all styles are now handled in an external css stylesheet. 3 stylesheets currently available:
- classic.css: follows the style of previous versions of the plugin
- icinga.css: follows the style of the icinga classic ui (default)
- nagiosxi.css: follows the style of the nagiosxi ui
- change: moved javascript to external file & reviewed to not generate conflicts with styles defined in css stylesheets
- feature: new interface inclusion/exclusion system distinguing properties and traffic load. 3 levels of inclusion/exclusion:
- global (exclude/include)
- globally include/exclude interfaces to be monitored
- excluded interfaces are represented by black overlayed rows in the interface table
- by default, all the interfaces are included in this tracking. Excluding an interface from that tracking is usually done for the interfaces that we don't want any tracking (e.g. loopback interfaces)
- traffic tracking (exclude-traffic/include-traffic)
- include/exclude interfaces from traffic tracking
- traffic tracking consists in a check of the bandwidth usage of the interface,and the error/discard packets.
- excluded interfaces are represented by a dark grey (css dependent) cell style in the interface table
- by default, all the interfaces are included in this tracking. Excluding an interface from that tracking is usually done for the interfaces known as problematic (high traffic load) and consequently for which we don't want load tracking
- property tracking (exclude-property/include-property)
- include/exclude interfaces from property tracking.
- property tracking consists in the check of any changes in the properties of an interface, properties specified via the --track-property option.
- excluded interfaces are represented by a dark grey (css dependent) cell style in the interface table
- by default, only the "operstatus" property is tracked. For the operstatus property, the exclusion of an interface is usually done when the interface can be down for normal reasons (ex: interfaces connected to printers sometime in standby mode)
- feature: new "Action" column, which currently contains a graph icon link, replacement of the "link on row" in previous versions of the plugin. This field allows other actions to be added in future.
- feature: installer (./configure; make; makeinstall...) for an easier install. It includes:
- the check of all required softwares / perl libraries
- automatic sudo and apache/httpd configuration
- a testing mode
- change: extracted snmp related and more general functions from the scripts and created libraries of functions, for better maintenance and to be used in other future plugins
- change: commonly used options are now default.
- change: detailed output is now default (use of outputshort options to reduce it).
- bugfix: fixed redondant short option, forcing conversion of oper/admin status from integer to string
- feature: error/discard packets
- checked against thresholds (rates)
- presented via a specific column in the interface table. Cell background color changes in case of warning/critical state
- change: new format for warning & critical commandline options for more simplicity:
--warning/--critical ,,
- unique threshold for in/out statistics (ex: loadin/loadout)
- one option for all traffic checks (load, pkt errors, pkt discards) to avoid too many options
ex: --warning 50,0,0 --critical 80,100,100
- feature: snmp v2c & v3 support
- feature: 64bit counters support (octets in/out & speed) for interfaces >=10Gbps
- feature: config file option --config. A config file can be used to load a set of configuration variables.
- feature: new property tracked ifDuplexStatus (duplex mode) (idea from Oliver Windt)
- feature: index page for interface table selection
- feature: new option perfdataformat which permit to choose the perfdata types to return
- full (default): generated performance data include plugin related stats, interface status, interface load stats, and packet error stats
- loadonly: generated performance data include plugin related stats, interface status, and interface load stats
- globalonly: generated performance data include only plugin related stats
'loadonly' should be used in case of too many interfaces and consequently too much performance data which cannot fit in the nagios plugin output buffer and are truncated. By default, its size is 8k and can be extended by modifying MAX_PLUGIN_OUTPUT_LENGTH in the nagios sources (icinga.h/nagios/h).
- bugfix: added missing ifoperstatus possible values (reported by Michael Gumbinger)
- enhancement: operational status are not tracked when changing from "xxx" to "up"
- enhancement: detect incoherent load > if speed (in case of node restart, switch from v1 to v2c,...) and avoid bad performance data
- experimental feature: new --(no)unixsnmp option to switch between full use of perl's netsnmp and use of Unix snmp utils. Need more testing in real environment, so still EXPERIMENTAL.
v0.02
- enhancement: cleaner and step-by-step output when resetting stats
- feature: finished option 'track', replacement of nochange, which list the field concerned by tracking. Default: only ifOperStatus.
- feature: added operstatus as performance data for interface status tracking
- enhancement: don't alert for interfaces OperStatus going to "up"
- bugfix: avoid global perfdata to be returned in case of initial run or no history, as the perfdata are wrong and produce drops on graphs
- enhancement: rewrite the plugin to follow the nagios plugin development guidelines: use of "use warnings"
- enhancement: added excludeportperf and includeportperf, allowing to filter which tracked interfaces generates performance data
- bugfix: port performance data now depend on what data are available. ex: unrouted vlans doesn't have ifOctetIn statistics on catalyst
- bugfix: corrected and completed physical interface pattern matching in check_for_unused_interfaces
- feature: option "cisco" for specific cisco info requests and printing
- feature: 2 new output cases, reflecting the leak of data for calculation. For initial run and for no valid historical dataset
- feature: operstatus in pnp4nagios port templates, for easier interface status tracking.
- enhancement: normal plugin output includes now load thresholds
- feature: --grapher and --grapherurl options to specify which grapher solution is used. And alternative url can be specified.
- feature: --short or --long options for less or more detailed plugin outputs. Using --long allow to clearly identify problematic interfaces directly in the output of the plugin
- enhancement: some options can now be specified multiple times and are concatenated. Exemple: --exclude 'pat1' --exclude 'pat2,pat3'. Concerned options: track, exclude, include, excludeportperf, includeportperf
v0.01
- enhancement: better pnp4nagios integration
- enhancement: better documentation on usage
- enhancement: debug mode review
- feature: added bits output feature for performance data
- enhancement: check_multi perfdata output type with interface name. Global and Port perfdata types
- feature: pnp6 templates
- enhancement: restricted the html link generated as the pluging output to "[detail]", for better integration in third party tools such as nagstamon,...
![[toggle]](javascript:icntoggle('mod-shoutboxr2','module.png'); "[toggle]"){kind=small}