Welcome to docs.opsview.com

Opsview Community is deprecated - read more here. Showing the most current documentation for Opsview Core

Changes

This page lists the major changes between releases of Opsview Core, our open source product.

Opsview Pro/Enterprise

Upgrade from Opsview Core to our commercial products, Opsview Pro or Enterprise, for extra features including:

  • Dashboard
  • Auto-Discovery
  • Distributed monitoring, including slave server clustering
  • NetFlow network analytics

If you are using a version of Opsview Community (3.20120424 or earlier) and want to upgrade to Opsview Core, then please review our upgrade documentation before attempting to upgrade your system.

Opsview Core 20131016

Enhancements

  • Debian 7 Wheezy officially supported as new platform
  • Updated Nagios code to Sept 2013
  • Default SNMP service checks support SNMPv3 by default
  • Set the maxiumum number of web requests an opsview_web_server daemon will process before restarting
  • Add ssl option to check_apache_performance
  • Include keywords in email notifications
  • opsview-agent init scripts updated to include 'status' method
  • Changed default host check interval to 5 minutes for new hosts so hosts are now regularly checked
  • Hosts without a host check command get assigned a dummy check instead
  • Reduced ndo.dat file sizes by removing duplicated information for LONGOUTPUT

Notices

  • The Public role has the VIEWPORTACCESS access only. Saving this role now removes all other accesses
  • Checks using ampersands (&) in arguments will give different results with the latest Nagios. This is likely to be incorrectly specified arguments. For instance, check_http -u /page?arg=1&arg=2 should quote special characters: check_http -u '/page?arg=1&arg=2'
  • The Nagios CGIs now have escape_html_tags=1 by default, so HTML will not be displayed. This matches with Opsview's status screens

Fixes

  • Added missing notification and event handler macros: HOSTDOWNTIME, SERVICEDOWNTIME
  • opsview-web init script now kills children as well as the parent, to ensure clean shutdown
  • check_opsview_ndo_import misses out .ignored files from being counted as age of NDO files
  • Fixed incorrect caching of admin edit pages
  • Renamed 'invert plugin results' command 'negate' to prevent clashes when used directly
  • Fixed multiple MySQL performance checks clashing on the same host and giving UNKNOWN results
  • Fixed refresh button on status.cgi page
  • Fixed freshness checking using refresh_state plugin
  • Fixed tab characters in plugin output not appearing in Opsview status screens

Opsview Core 20130903

Fixes

  • Fixed perl NDO updates where updates would hang if files were of a specific size
  • Fixed opsview_master_housekeep complaining about a missing subroutine
  • Fixed warnings when searching events by datetime objects

Opsview Core 20130822

Enhancements

  • Added new OS - Linux SNMP opspack for monitoring Linux hosts via SNMP
  • Speed up reload configuration generation by 25% when using lots of attributes or SNMP interfaces
  • Added AQL as a method of sending notifications
  • Improved Runtime database updates so post-reload actions are done in transactions and reduced SQL calls by 50%
  • Improved role pages with better description of access
  • Access control additions:
    • New CONFIGURECONTACTS, CONFIGUREROLES, CONFIGUREHOSTGROUPS. On upgrade, any role with CONFIGUREVIEW access will inherit these
    • New NAGVIS access. On upgrade, any role with VIEWALL and VIEWSOME will include NAGVIS for backwards compatibility. Recommend restriction to VIEWSOME
  • Updated ACL checking for consistency
  • Improved Debian postinstallation script so that only prompts for mysql root password on new installs
  • Random passwords created for database connections for new installs
  • Add in a local shell profile file that will not be lost on an upgrade
  • Improved CSS for admin pages to line up checkboxes, radio buttons and text
  • Improved NSCA daemon syslog entries to include client IP addresses
  • Added in missing option aliases to nlcat
  • Improved “Stop Alert After” so that state changes will not be blocked
  • Patch from Nagios Core to improve error message when plugin times out
  • Improved SNMP interface speed detection so that speeds that are 0 in ifXSpeed area of SNMP will use ifSpeed if possible
  • Plugin output has HTML tags escaped in the Hostgroup Hierarchy pages (use markdown instead if it is required)
  • Plugin output has newlines converted to <BR>'s in the Hostgroup Hierarchy pages
  • Changed Opsview Agent/NRPE's pid file directory to /var/tmp to support Solaris zones
  • Improved Opspack install process to always copy plugins and retry next time for conflicts
  • Ensure ospview-agent, opsview and opsview-web init.d scripts conform much better to standards
  • Solaris agent fixes:
  • - Use SMF instead of /etc/init.d/ script to start agent
  • - Shut down agent on package removal
  • - Compile all binaries to correct 32 or 64 bitness instead of mixed
  • - Remove NSCA binary to reduce 2rd part dependancies (since it is rarely used)
  • - Remove in place upgrade functionality as it wasn't reliable
  • - Fix warning on package install for registering FS space usage

Notices

  • On Ubuntu, MySQL must be running for Opsview to start up on a reboot. Opsview startup waits until MySQL is ready
  • Nagvis is now blocked by a new access, but configuration is required at Apache to restrict when upgrading
  • There are new service checks to monitor Opsview's housekeeping functions. These need to be manually added For further information. See documentation at http://docs.opsview.com/doku.php?id=opsview-core:upgrading

Fixes

  • Fixed XSS and CSRF exposure - Thanks to Charlie Eriksen via Secunia SVCRP. Fixes CVE-2013-5694 and CVE-2013-5695
  • Fixed CONTACTGROUPLIST environment macro, broken in Nagios 4 upgrade
  • Re-applied Nagios plugin limit of 16K
  • Re-enable contact notification information into Runtime
  • New notification methods will be set to run on master when upgrading to Opsview Pro/Enterprise
  • Fixed multiple “Starting Opsview Web” messages in opsview-web.log
  • check_snmp_interfaces_cascade now displays an informational message if there are no SNMP tables for extended throughput data
  • Fixed setting of sticky acknowledgements in mass acknowledgements page
  • Fixed REST API /runtime/service for VIEWSOME users
  • Added missing notification macros: HOSTDURATION, HOSTNOTIFICATIONNUMBER, NOTIFICATIONAUTHOR, NOTIFICATIONCOMMENT, SERVICEDURATION, SERVICENOTIFICATIONNUMBER
  • Fixed menu Network outages link only appearing if VIEWALL is granted for user
  • Fixed status pages where a long host group hierarchy was being hidden from view
  • Fixed contextual menus on performance style viewport when Opsview Web rebased
  • Fixed Safari bug where changing passwords were not validating correctly
  • Fixed some embedded newlines not being interpreted correctly for plugin output in email notifications
  • Fixed demo user allowing host interfaces to be changed without CONFIGURESAVE permission

Opsview Core 20130522

Enhancements

  • Notification profiles and shared notification profiles now have a field to prevent the sending of alerts after a defined number
  • Tool to help convert Nagios 3 command arguments to Nagios 4 style. See documentation
  • REST API has a convertuoms=1 parameter to convert performance data uoms into their base values (eg, MB to bytes, ms to seconds)
  • Allow changing Starman server options from opsview_web.yml file. Defaults are now closer to previous HTTP::Prefork behaviour
  • SNMP throughput thresholds now have an expanded syntax to handle input and output throughput separately as well as ranges
  • Remove some more specific words from SNMP interface ifDescr to reduce instances of duplicate interface ids being used (level 4 added)
  • Acknowledging a problem via the host or service state information cgi page now uses the standard Opsview form
  • check_sql_advanced query duration also takes account of fetching the rows. Also amend execution duration calculation to be seconds, not milliseconds and always include it in perfdata
  • Any new Opspacks will be installed upon future upgrades
  • New Opspack: MySQL Server, replacing existing MySQL
  • New Opspack: Apache HTTP Server, replacing existing Apache HTTP
  • Management URLs open in a new window for http and https methods
  • check_snmp_interfaces_cascade now reports timings on main execution parts and wraps cache updates in a single transaction. Previous time metric is now renamed to totaltime
  • Downgrade check_snmp_interfaces_cascade errors with invalid percentage utilisation (critical) and long plugin execution time (warning) to be unknowns instead

Notices

  • Removed OS - Linux Advanced opspack from fresh installs

Fixes

  • Fixed Nagios 4 return code 7 issues and environment variables not being consistently applied in notifications
  • Fixed iOS host notifications not showing correct detail
  • Fixed validation so that spaces are not allowed at beginning of service check names
  • Fixed check_snmp_ifstatus as wrong snmp-version specified
  • Possible fix for check_snmp_interfaces with over 100% interface utilisation
  • Fixed account menu being displayed under menu options in IE8
  • Merge of upstream Nagios 4 fixes: backticks checked within double quotes, close parenthesis detected, shell variables and subcommmands detected, closing all file descriptors in workers, various other fixes
  • Fixed semicolons in command definitions not being interpreted as comments
  • Fixed ~ as needing to be executed by shell
  • Fixed Debian nagios-nrpe-server package not being a conflict
  • Fixed check_snmp_ifstatus plugin to cope with gaps in nodes when polling SNMP interfaces
  • Fixed SNMP polling of device interfaces with no ifDescr set
  • Fixed paths to images in events view and dashboard splash page for a rehomed web app
  • Fixed adding custom contact variables from notification methods into environment variables for notification scripts to use
  • Fixed notify_by_rss's maximum number of items, which was off by one
  • Added missing notification macros: LASTHOSTDOWN, LASTHOSTCHECK, LASTHOSTUNREACHABLE, LASTHOSTUP, LASTSERVICECHECK, LASTSERVICEOK, LASTSERVICEWARNING, LASTSTATECHANGE
  • Fixed attributes with numbers in the name when processing commands
  • Redirected 'call_nmis' cronjob output to the existing nmis.log logfile
  • Stop running MRTG checks on a reload when SNMP is not enabled
  • Fixed check_snmp_interfaces_cascade where no ifSpeed is matched to an interface
  • Fixed issue where Host Attributes Value of 0 being converted to empty string in UI

Opsview Core 20130304

Enhancements

  • Updated NRPE to 2.14 due to potential security exposure
  • check_snmp_sysinfo now returns CRITICAL if it cannot connect to SNMP
  • Allow authtkt cookie to be set when logging into Opsview REST API
  • Allow tracing of individual hosts in check_snmp_interfaces_cascade
  • Restore clicking on product name in footer to get build version information

Notices

  • Due to Nagios 4's different parsing rules, service checks with arguments that contain a backslash (\) should be checked to see if they still work correctly
  • NRPE no longer allows shell metacharacters to be passed as command line arguments to the remote plugin. If attempted, this will return “CHECK_NRPE: Received 0 bytes from daemon. Check the remote server logs for error messages”
  • You may find that Interface checks return ”… has an interface speed of 0, so cannot check a percentage threshold” - you will need to change the thresholds to be an absolute value, instead of a percent value

Fixes

  • Fixed incorrect SNMP version definition and timeout values
  • Fixed check_snmp_weblogic_jmsmsgbridge to reincorporate caching host data to a file
  • Fixed double escape of backslashes, which can affect Windows checks
  • Fixed Nagios CGI for hostnames with funny characters
  • Fixed SNMP - MIB-II opspack so that a dependency is set for active polled SNMP checks to have dependency on the agent
  • Fixed sensitive arguments field which was missing from UI in error
  • Fixed Nagios 4 execution of event handlers
  • Possible fix for log file archive rotation overwriting - check for “Warning: Another log rotation was requested for…” messages
  • Plugin output with linefeeds are now read correctly
  • Command arguments with double back slashes, eg \\, are now not double escaped
  • Fixed the profile being updated when using bash

Opsview Core 20130212

Features

  • Upgraded to latest Nagios 4 pre-release version. Reduces system load by 60%
  • Replaced ndo2db with perl based importer. Reduces time taken to import by 40%
  • Improve configuration generation time by an average of 30% for systems with complex service dependencies, systems with complex parent/child relationships or systems with many host attributes. Includes a reduction in SQL calls by up to 35%

Platforms

  • IE10 on Windows 8 is now a supported browser
  • IE7 is deprecated
  • Solaris 10 is not included in this release due to issues with Nagios 4. Solaris is a deprecated platform

Enhancements

  • Added enhanced monitoring for Oracle/PostgreSQL (new service checks to existing Host template)
  • Attributes are now allowed to use numbers in the name
  • /rest/runtime/network?fromhostname=host now returns list back roughly based on depth of host
  • Search box in the main interface now also searches host descriptions
  • Amend keyword references from 'keyword_N' to 'kN_<KEYWORD>' within the configuration files
  • Added 'negate' and 'urlize' to list of available servicecheck plugins
  • Allow for some default options for hosts and servicechecks to be amended in opsview.conf
  • Remove unnecessary contacts from generated configuration (US254)
  • Libmcrypt is now included in the Opsview repository for RHEL6, to remove dependency on rpmforge
  • Switch Downtime link to goto Opsview page, which collapses related downtimes together
  • Added (disabled) HTTP header expiry options for apache into the example proxy configuration file
  • It is now possible to specify an SNMP port greater than 32768
  • Added logging for check_snmp_interfaces_cascade for situations where threshold > 100%
  • Changed Nagios CGI date format to be iso8601 on fresh installs
  • Updated NRD configuration to apply results into checkresults with timestamp value for Nagios to process in order
  • Removed all use_prefork and use_lighttpd references from code
  • Conversion routines between bytes and megabytes now use 1000^2, rather than 1024^2. This may affect the graphing display although all historic data is held in a raw format
  • Updated half of SNMP plugins to support SNMPv3

Notices

  • Only a subset of Nagios environment macros are available. Plugins will only get environment variables if specified within the plugin. Notifications and event handlers have a limited set of macros. See http://docs.opsview.com/doku.php?id=opsview-core:notificationmethods for a list
  • REST error messages have colons removed
  • Updated primary keys for nagios_hostchecks and snmptrapexceptions to BIGINT - those tables will be truncated on upgrade
  • Some check_snmp_* plugins were referencing livehosttable.db to find an SNMP community value. This functionality has been removed

Fixes

  • Fixed check_snmp_interfaces_cascade so that interfaces with zero speed (eg VLANs) have an appropriate error message when trying to do a percentage threshold check
  • Fixed HH displaying breadcrumbs for host groups that user did not have permission for
  • Fixed apache configuration for redundancy, specifically removing apache_proxy_ssl.conf and replacing with apache_ssl.conf
  • Fixed messages for set state
  • Fixed display of check_snmp_interfaces_cascade so that bits/s to Gbps is using division by 1000^3 rather than 1024^3
  • Fixed rc.opsview-agent init script to pick up when nrpe is running on an incorrectly recorded pid
  • Fixed installation on SLES11 where check_http_authentication would cause upgrades to hang
  • Fixed SOAP::Lite perl module error
  • Workaround for librrd bug which hangs the RRD import. Now import will log a message to opsviewd.log after 10 seconds and continue
  • Change to using Cache::BerkelyDB instead of deprecated Cache::FastMmap
  • Fixed opsview_cronjobs to append to log file rather than overwrite
  • Added RPM conflicts so opsview-slave cannot install with opsview-core
  • Unreachable hosts now show in cornflour colour
  • The Keyword Summary page would show a keyword as critical instead of unreachable, if the host was unreachable but all services OK
  • In IE9, context menus in some places would disappear when another context menu item was selected
  • The check_route plugin handles errors
  • Limited the Keywords Summary to those for which access is granted
  • Our mod_auth_tkt RPM now obsoletes the standard one. Previously, it would conflict.
  • The web cache (/opt/opsview/work/web-cache) is now cleared upon starting opsview-web
  • Fix cell borders in keyword view
  • The reload timings are now centred
  • Fixed acknowledgement option to only appear in menus if the object is unhandled
  • Fixed possible error message on RH systems when using init scripts
  • Fixed bug with community string in check_snmp_netscreen_activesessions and check_snmp_netscreen_cpu
  • Fixed resetting of UUID when known bad UUIDs detected
  • Fixed adding opsview profile to both .profile and .bashrc files
  • Fixed how status.cgi stats are calculated, which means audible alerts work correctly again from Nagios CGIs

Opsview Core 20120925

Features

  • Added push notifications for IOS devices

Enhancements

  • REST API for listing currently running monitoring servers
  • REST API for /status/hostgroup can now return information about service groups
  • REST API for /status/hostgroup allows searching from a point in the heirarchy downwards using fromhostgroupid
  • Remove some specified words from device interface ifDescr information to reduce instances of duplicate interface
  • Warnings on multiple hosts using same IP when hosts are configured for MRTG
  • REST API for /runtime/network can now accept maximum limit of entries returned and provides counts information
  • REST API for /runtime/network can now accept a parent host for topology map

Notices

  • Removed runtime.opsview_monitoringclusters, which was not used
  • REST API for /runtime/network result is now wrapped as list element, next to counts information

Fixes

  • Fixed possible issues when removing Debian packages
  • Fixed opsview_sync_ldap where apostrophes in a user's name
  • Fixed issue where interfaces inheriting default threshold values did not have the threshold set correctly, when using check_snmp_interfaces_cascade
  • 'Edit host configuration' context menu link has been fixed to use the CONFIGUREHOSTS access
  • Link to the auditlog on reload page no longer appears upon successful reload
  • Fixed issue with non-authenticated user requesting access limited user menus

Opsview Core 20120809

Features

  • Shared notification profiles with a new access point of CONFIGUREPROFILES
  • Support for Ubuntu Linux 12.04

Enhancements

  • Allow to exclude handled hosts and services when calculating overall state in the keywords summary page
  • Allow for generating URLs in notifications to link to object information pages
  • Store downtimes and acknowledgements in statehistory table. Upgrades may take a few minutes for schema changes
  • REST API for events now contains additional acknowledgement and downtime information
  • Small improvement to ndo2db when querying for objects
  • New util: list_timeperiod_users
  • New REST API parameter when searching performance metrics: collapse_multiple_services, so multiple services are only returned once
  • REST API for performance data now returns summarised information on 95th percentile and standard deviation
  • REST API now returns the server's timezone offset based on current time
  • Allow ndo.dat to be moved more easily
  • You can no longer edit a contact's username through the web UI, as this affects 3rd party systems (dashboard, reporting module, Nagvis)
  • Updated Nagios::Plugin and dependencies
  • Updated Nagios Plugins to 1.4.17 (latest nightly)
  • Log the full query_host command into opsview-web.log if there is an error picked up
  • Updated Perl modules as of July 2012
  • Allow check_snmp performance data units to be overridden

Notices

  • Remove support for Ubuntu Hardy
  • Opsview Web does not install new monitor packs on start - use install_all_opspacks command line utility
  • Renamed install_monitor_pack to install_opspack and create_monitor_pack to create_opspack
  • Monitoring packs are now installed from var/spool/opspacks (was var/spool/monitorpacks)

Fixes

  • Fixed DB issue for REST API audit logs when opsview db renamed
  • Correct Solaris OpenCSW package dependencies
  • Fixed Solaris opsview-base postinstall error 'die: not found'
  • Fixed dependency of snmpget on RHEL
  • Fixed time jump logic to take account of the threshold for both forward and reverse jumps
  • Fixed nightly backup attempt of lost+found in /usr/local/nagios
  • Fixed fresh debian install with Nagvis permissions for www-data
  • Fixed Nagvis configuration issues because of Opsview banner
  • Start URL (start_url in opsview_web.yml) is now taken as-is allowing query string params. NOTE: If you rebase Opsview Web to a different location, you will need to amend this start_url appropriately
  • Fixed audible alerts in events view
  • Fixed net-snmp-perl requirement on RedHat
  • Ensure nagcmd group is added to Apache user
  • Allow check_snmp_perfstats to use SNMP v2c
  • Fixed SLES11 dependency on apache2-mod_php5
  • Fixed opsview not always correctly starting on reboot (Thanks to Mark Addiss)
  • Added javascript-common as conflicting package on Debian/Ubuntu
  • Fixed opspacks not being installed correctly when no plugins are included
  • Fixed check_apache_performance where metric returns 0 as value

Opsview Core 20120516

Features

  • New RESTAPI to return current performance data

Enhancements

  • NRPE now only accepts AES256 or AES128 ciphers, thus reducing security exposure from clients accepting lower levels of encryption
  • NRD no longer uses a PID file
  • Use prefork by default for Opsview Web
  • New TESTALL and TESTSOME access points for running service checks dynamically. UI changes in next release
  • New TESTCHANGE access point to allow changing of arguments to plugin for testing Upgrade rules: ACTIONALL will get TESTALL, ACTIONSOME → TESTSOME, ADMINACCESS → TESTCHANGE
  • Removed seldom used /admin/notifications page

Fixes

  • Fixed MRTG collection so that it uses system level perl libraries
  • Fixed host icons when upgrading from previous versions of Opsview
  • Fixed upgrades when old libraries are loaded by opsview-web
  • Fixed postinstall on RPM and Solaris so that creates same template opsview.conf as Debian
  • Fixed REST API status call so that state_type input can be entered as soft, hard, 0 or 1
  • Fixed snmpwalk from Opsview service check edit screen on systems that are missing MIBs by default
Navigation
Print/export
Toolbox