Check_MK

Check_MK is an open source, multiplatform and freely distributed web-based application designed from the offset to provide users with a new way to collect data from operating systems or network components.

The software obsoletes the NRPE, check_by_ssh, NSClient and check_snmp applications. It provides significant reduction of CPU usage on Nagios hosts, as well as automatic inventory of items to be checked on hosts.

The application includes three awesome tools, Multisite, a full featured status GUI (Graphical User Interface) for Nagios, based on Livestatus; MK Livestatus, a Nagios event broker module; and WATO, a configuration GUI (Graphical User Interface).

Features at a glance

Key features include a beautiful and attractive method for configuring the Nagios monitoring system by activating configuration generator and automatic service recognition instead of using Nagios’ config data, support for contracting each host once per check interval, as well as support for sending check results to Nagios as passive checks, saving a lot of resources on both client and server systems.

Additionally, Check_MK offers a flexible and simple notification system that can be easily configured. The notification system supports defining and configuration of multiple channels for different users, but users can also configure the notification system themselves.

Supports mobile devices

Another interesting features is the ability to use the application straight from a mobile device, such as smartphone or tablet. It allows users of mobile devices to easily access status data, as well as to execute “Set for Downtimes” and “Acknowledge” commands. All mobile devices will be automatically recognized.

What is new in this release:

• Core & Setup:
• 1035 FIX: Do not fail on errors in *.mk files anymore - except in interactive mode...
• Checks & Agents:
• 0994 FIX: agent plugin smart: fixed syntax error
• 0989 FIX: logwatch.ec: Fix forwarding multiple messages via syslog/TCP...
• 0943 FIX: if.include: fixed incorrect traffic percentage values in the check output of if checks...
• 1037 FIX: bintec_info: support bintec RXL12500
• 0827 FIX: lnx_thermal: Not checking active trip points (e.g. cooling device triggers) anymore
• Multisite:
• 1036 FIX: doc/treasures/downtime: fix --url option, better error output
• WATO:
• 0990 FIX: Fix HTTP error handling in bulk inventory...
• 1004 FIX: Fix exception when saving rules, caused by empty item
• Event Console:
• 1040 FIX: Avoid sporadic errors when checking event state in Event Console...
• Livestatus:
• 0988 FIX: livedump: Fix exception in case no contact groups are defined for a service

What is new in version 1.2.4:

• This release comes more than one thousand changes.
• Check_MK Micro Core was introduced, boosting performance of huge installations a lot.
• Multisite has been extended with an availability reporting module.
• The Business Intelligence rules can be edited via a Web browser using WATO.
• Many new checks were added, including ESX and SAP R/3 monitoring, and there were several usability and performance improvements.

What is new in version 1.2.2p3:

• 0041 FIX : setup.py now handles non existing wwwuser gracefully...
• 0040 : Add agent plugin to test local hostname resolving...
• 0020 FIX : Inventory problem with inventory_processes parameter...
• 0000 : Improved performance of LDAP sync by refactoring the group sync code
• 0042 FIX : Removed debug outputs from service inventory...

What is new in version 1.2.3i2:

• Core:
• New option -B for just generating the configuration
• Introduced persistent host address lookup cache to prevent issues loading an unchanged configuration after a single address is not resolvable anymore
• Assigning a service to a cluster host no longer requires a reinventory
• Setting a check_type or service to ignore no longer requires a reinventory Note: If the ignore rule is removed the services will reappear
• Config creation: The ignore services rule now also applies to custom, active and legacy checks
• Predictive monitoring: correctly handle spaces in variable names (thanks to Karl Golland)
• New man page browser for console (cmk -m)
• New option explicit_snmp_communities to override rule based SNMP settings
• Preparations for significant SNMP monitoring performance improvement (It's named Inline SNMP, which is available as special feature via subscriptions)
• Allow to specify custom host check via WATO (arbitrary command line)
• Implement DNS caching. This can be disabled with use_dns_cache = False
• Livestatus:
• new service column staleness: indicator for outdated service checks
• new host column staleness: indicator for outdated host checks
• Checks & Agents:
• esx_hostystem multipath: criticize standby paths only if not equal to active paths
• mk_logwatch: fixed bug when rewriting logwatch messages
• check_mk: Re-inventory is no longer required when a service is ignored via rule
• check_mk: Now possible to assign services to clusters without the need to reinventorize
• lnx_if: Fixed crash on missing "Address" field
• viprinet_router: Now able to set required target state via rule
• windows_agent: Now available as 64 bit version
• agent_vsphere: fix problem where sensors were missing when you queried multiple host systems via vCenter
• cached checks: no longer output cached data if the age of the cache file is twice the maximum cache age
• windows agent: no longer tries to execute directories
• fileinfo: no longer inventorize missing files(reported by windows agent)
• New checks for Brocade fans, temperature and power supplies
• cluster hosts: removed agent version output from Check_MK service (this was misleading for different agent versions on multiple nodes)
• job check: better handling of unexpected agent output
• lnx_thermal: Added check for linux thermal sensors (e.g. acpi)
• hwg_temp: Make WATO-Rule "Room Temperature" match, add man page, graph and Perf-O-Meter
• ps.perf: Support Windows with new plugin "psperf.bat". wmicchecks.bat is obsolete now.
• Special Agent vSphere: support ESX 4.1 (thanks to Mirko Witt)
• esx_vsphere_object: make check state configurable
• mk_logwatch: support continuation lines with 'A'. Please refer to docu.
• mk_oracle: Added plugin for solaris
• win_netstat: New check for Windows for checking the existance of a UDP/TCP connection or listener
• ps/ps.perf: allow to set levels on CPU util, optional averaging of CPU
• diskstat: Agent is now also processing data of mmcblk devices
• qmail: Added check for mailqueue
• cisco_locif: removed obsolete and already disabled check completely
• fc_brocade_port: removed obsolete check
• fc_brocade_port_detailed: removed obsolete check
• tsm_stgpool: removed orphaned check
• vmware_state: removed ancient, now orphaned check. Use vsphere_agent instead.
• vms_{df,md,netif,sys}: remove orphaned checks that are not needed by the current agent
• tsm: Added new TSM checks with a simple windows agent plugin
• windows_agent: now starts local/plugin scripts in separate threads/processes new script parameters cache_age, retry_count, timeout new script caching options "off", "async", "sync"
• windows_agent: increased maximum local/plugin script output length to 512kB (output buffer now grows dynamically)
• jolokia_metrics: fixed incorrect plugin output for high warn/crit levels
• jolokia_metrics.uptime: Added pnp template
• hyperv: Added a check for checking state changes.
• df / esx_vsphere_datastore: now able to set absolute levels and levels depending on total disk space of used and free space
• cisco_wlc: New check for monitoring cisco wireless lan access points
• cisco_wlc_clients: New check for the nummber of clients in a wlc wifi
• df: Negative integer levels for MB left on a device
• win_printers: Monitoring of printer queue on a windows printserver
• cisco_qos: Updated to be able to mintor IOS XR 4.2.1 (on a ASR9K device)
• New active check, check_form_submit, to submit HTML forms and check the resulting page
• mk-job: /var/lib/check_mk_agent/job directory is now created with mode 1777 so mk-job can be used by unprivileged users too
• ADD: etherbox: new check for etherbox (messpc) sensors. currently supported: temperature, humidity, switch contact and smoke sensors
• cisco_wlc_client: now supports low/high warn and crit levels
• cisco_wlc: now supports configuration options for missing AP
• agent_vsphere: completely rewritten, now considerably faster vCenter is still queried by old version
• windows_agent: windows eventlog informational/audit logs now reported with O prefix
• mk_logwatch: ignored loglines now reported with an "." prefix (if required)
• apache_status: Nopw also supports multithreaded mpm
• windows_agent: now able to suppress context messages in windows eventlogs
• agent_vsphere: completely rewritten, now considerably faster vCenter is still queried by old version
• windows_agent: windows eventlog informational/audit logs now reported with O prefix
• mk_logwatch: ignored loglines now reported with an "." prefix (if required)
• check_mk-if.pnp: fixed bug with pnp template on esx hosts without perfdata
• jolokia checks (JVM): uptime, threads, sessions, requests, queue now configurable via WATO
• vSphere checks: secret is not shown to the user via WATO anymore
• WATO rule to check state of physical switch (currently used by etherbox check)
• cisco_wlc: Allows to configure handling of missing AP
• logwatch.ec: show logfiles from that we forwarded messages
• FIX: blade_blades: Fixed output of "(UNKNOWN)" even if state is OK
• FIX: apache_status: fix exception if parameter is None
• FIX: hr_mem: handle virtual memory correct on some devices
• FIX: apache_status agent plugin: now also works, if prog name contains slashes
• FIX: check_dns: parameter -A does not get an additional string
• FIX: cisco_qos: Catch policies without post/drop byte information
• FIX: cisco_qos: Catch policies without individual bandwidth limits
• FIX: windows_agent: fixed bug on merging plugin output buffers
• FIX: esx_vsphere_datastores: Fix incomplete performance data and Perf-O-Meter
• FIX: cleaned up fileinfo.groups pattern handling, manual configuration is now possible using WATO
• FIX: check_mk-ipmi.php: PNP template now displays correct units as delivered by the check plugin
• FIX: check_disk_smb: Remove $ from share when creating service description. Otherwise Nagios will not accept the service description.
• FIX: mrpe: gracefully handle invalid exit code of plugin
• Notifications:
• notify.py: Matching service level: Use the hosts service level if a service has no service level set
• notify.py: fixed bug with local notification spooling
• HTML notifications: Now adding optional links to host- and service names when second argument notification script is configured to the base url of the monitoring installation (e.g. http://// in case of OMD setups)
• HTML notifications: Added time of state change
• Multisite:
• Finally good handling of F5 / browser reloads -> no page switching to start page anymore (at least in modern browsers)
• User accounts can now be locked after a specified amount of auth failures (lock_on_logon_failures can be set to a number of tries)
• Column Perf-O-Meter is now sortable: it sorts after the *first* performance value. This might not always be the one you like, but its far better than nothing.
• logwatch: Logwatch icon no longer uses notes_url
• Inventory screen: Host inventory also displays its clustered services
• Rules: Renamed "Ignored services" to "Disabled services" Renamed "Ignored checks" to "Disabled checks"
• Sorter Host IP address: fixed sorting, no longer uses str compare on ip
• Views: New: Draw rule editor icon in multisite views (default off) Can be activated in global settings
• New global multisite options: Adhoc downtime with duration and comment Display current date in dashboard
• LDAP: Using asynchronous searches / added optional support for paginated searches (Can be enabled in connection settings)
• LDAP: It is now possible to provide multiple failover servers, which are tried when the primary ldap server fails
• LDAP: Supporting posixGroup with memberUid as member attribute
• LDAP: Added filter_group option to user configuration to make the synchonized users filterable by group memberships in directories without memberof attributes
• LDAP: Moved configuration to dedicated page which also provides some testing mechanisms for the configuration
• Added option to enable browser scrollbar to the multisite sidebar (only via "sidebar_show_scrollbar = True" in multisite.mk
• Added option to disable automatic userdb synchronizations in multisite
• Implemented search forms for most data tables
• New icons in view footers: export as CSV, export as JSON
• Availability: new columns for shortest, longest, average and count
• Editing localized strings (like the title) is now optional when cloning views or editing cloned views. If not edited, the views inherit the localized strings from their ancestors
• Added simple problems Dashboard
• New filter and column painter for current notification number (escalations)
• Added new painters for displaying host tags (list of tags, single tag groups). All those painters are sortable. Also added new filters for tags.
• Added painters, icon and filters for visualizing staleness information
• Improved filtering of the foldertree snapin by user permissions (when a user is only permitted on one child folder, the upper folder is removed from the hierarchy)
• "Unchecked Services" view now uses the staleness of services for filtering
• Globe dashlets make use of the parameter "id" to make it possible to provide unique ids in the render HTML code to the dashlets
• Multisite can now track wether or not a user is online, this need to be enabled e.g. via Global Settings in WATO (Save last access times of users)
• Added popup message notification system to make it possible to notify multisite users about various things. It is linked on WATO Users page at the moment. An image will appear for a user in the sidebar footer with the number of pending messages when there are pending messages for a user. To make the sidebar check for new messages on a regular base, you need to configure the interval of sidebar popup notification updates e.g. via WATO Global Settings.
• Event views: changed default horizon from 31 to 7 days
• New option for painting timestamp: as Unix Epoch time
• New filters: Host state type and Service state type
• FIX: better error message in case of exception in SNMP handling
• FIX: Inventory screen: Now shows custom checks
• FIX: Fixed locking problem of multisite pages related to user loading/saving
• FIX: Fixed wrong default settings of view filters in localized multisite
• FIX: line wrapping of logwatch entries
• FIX: Fixed button dragging bug when opening the view editor (at least in Firefox)
• WATO:
• Allow to configure check-/retry_interval in second precision
• Custom user attributes can now be managed using WATO
• Allow GIT to be used for change tracking (enable via global option)
• Hosts/Folders: SNMP communities can now be configured via the host and folders hierarchy. Those settings override the rule base config.
• Require unique alias names in between the following elements: Host/Service/Contact Groups, Timeperiods and Roles
• Removed "do not connect" option from site socket editor. Use the checkbox "Disable" to disable the site for multisite.
• Converted table of Event Console Rules to new implementation, make it sortable
• FIX: do validation of check items in rule editor
• FIX: More consistent handling of folderpath select in rule editor
• FIX: Now correctly handling depends_on_tags on page rendering for inherited values
• FIX: Changed several forms from GET to POST to prevent "Request-URI too large" error messages during submitting forms
• FIX: automation snmp scan now adhere rules for shoddy snmp devices which have no sys description
• FIX: Cisco ruleset "Cisco WLC WiFi client connections" has been generalized to "WLC WiFi client connections"
• FIX: Snapshot handling is a little more robust agains manually created files in snapshot directory now
• FIX: Slightly more transparent handling of syntax errors when loading rules.mk
• Notifications:
• Flexible Notification can now filter service levels
• FIX: check_tcp corrected order of parameters in definition
• Event Console:
• New global setting "force message archiving", converts the EC into a kind of syslog archive
• New built-in snmptrap server to directly receive snmp traps
• FIX: fix layout of filter for history action type
• FIX: better detect non-IP-number hosts in hostname translation

What is new in version 1.2.2p2:

• Core:
• FIX: better error message in case of exception in SNMP handling
• FIX: livecheck: fixed handling of one-line plugin outputs and missing \n (Thanks to Florent Peterschmitt)
• Checks & Agents:
• FIX: apache_status: use (also) apache_status.cfg instead of apache_status.conf
• FIX: jolokia_info: ignore ERROR instances
• FIX: f5_bigip_vserver: fix wrong OID (13 instead of 1), thanks to Miro Ramza
• FIX: f5_bigip_psu: handle more than first power supply, thanks to Miro Ramza
• FIX: ipmi_sensors: ignore sensors in state [NA] (not available)
• FIX: aix_lvm: handle agents that output an extra header line
• FIX: zfsget: do not assume that devices begin with /, but mountpoints
• FIX: ipmi_sensors: handle two cases for DELL correctly (thanks to Sebastian Talmon)
• FIX: check_dns: enable performance data
• FIX: free_ipmi: fix name of sensor cache file if hostname contains domain part
• FIX: ad_replication plugin: Fixed typo (Thanks to Dennis Honke)
• Multisite:
• List of views: Output the alias of a datasource instead of internal name
• FIX: fix column editor for join columns if "SERVICE:" is l10n'ed
• FIX: fix invalid request in livestatus query after reconnect
• WATO:
• FIX: convert editing of global setting to POST. This avoid URL-too-long when defining lots of Event Console actions
• FIX: LDAP configuration: allow DNs without DC=
• Event Console:
• FIX: fix icon in events check if host specification is by IP address
• Renamed "Delete Event" to "Archive Event" to clearify the meaning
• Notifications:
• FIX: contacts with notifications disabled no longer receive custom notifications, unless forced

What is new in version 1.2.3i1:

• Core:
• Agents can send data for other hosts "piggyback". This is being used by the vSphere and SAP plugins
• New variable host_check_commands, that allows the definition of an alternative host check command (without manually defining one)
• New variable snmp_check_interval which can be used to customize the check intervals of SNMP based checks
• setup: Added missing vars rrd_path and rrdcached_sock
• new variable check_mk_exit_status: allows to make Check_MK service OK, even if host in not reachable.
• set always_cleanup_autochecks to True per default now
• check_mk: new option --snmptranslate
• Multisite:
• New availability view for arbitrary host/service collections
• New option auth_by_http_header to use the value of a HTTP header variable for authentication (Useful in reverse proxy environments)
• New permission that is needed for seeing views that other users have defined (per default this is contained in all roles)
• New path back to the view after command exection with all checkboxes cleared
• Added plugins to config module to make registration of default values possible for addons like mkeventd - reset to default values works now correctly even for multisite related settings
• perfometer: Bit values now using base of 1000
• Added PNP tempate for check_disk_smb
• Dashboards can now be configured to be reloaded on resizing (automatically adds width/height url parameters)
• LDAP authentification: New config option "Do not use persistent connections to ldap server"
• Hosttags and auxiliary tags can now be grouped in topics
• Fixed output of time in view if server time differs from user time
• Event Console:
• New rule feature: automatically delete event after actions
• New filter for maximum services level (minimum already existed)
• New global setting: hostname translation (allows e.g. to drop domain name)
• New rule match: only apply rule within specified time period
• Checks & Agents:
• solaris_mem: New check for memory and swap for Solaris agent
• agent_vsphere: New VMWare ESX monitoring that uses pySphere and the VMWare API in order to get data very efficiently. Read (upcoming) documentation for details.
• new special agent agent_random for creating random monitoring data
• New checks: windows_intel_bonding / windows_broadcom_bonding
• Implemented SAP monitoring based on the agent plugin mk_sap. This must be run on a linux host. It connects via RFC calls to SAP R/3 systems to retrieve monitoring information for this or other machines.
• sap.dialog: Monitors SAP dialog statistics like the response time
• sap.value: Simply processes information provided by SAP to Nagios
• openvpn_clients: new check for OpenVPN connections
• if64_tplink: special new check for TP Link switches with broken SNMP output
• job: Monitoring states and performance indicators of any jobs on linux systems
• oracle_asm_diskgroups: Added missing agent plugin + asmcmd wrapper script
• oracle_jobs: New check to monitor oracle database job execution
• oracle_rman_backups: New check to monitor state of ORACLE RMAN backups
• jar_signature: New check to monitor wether or not a jar is signed and certificate is not expired
• cisco_qos: adhere qos-bandwidth policies
• check_disk_smb: WATO formalization for active check check_disk_smb
• if.include: new configurable parameters for assumed input and output speed
• cisco_qos: new param unit: switches between bit/byte display new param average: average the values over the given minute new params post/drop can be configured via int and float fixed incorrect worst state if different parameters exceed limit
• logwatch.ec: Added optional spooling to the check to prevent dataloss when processing of current lines needs more time than max execution time
• mounts: ignore multiple occurrances of the same device
• Linux agent: allow cached local/plugins checks (see docu)
• mem.include: Linux memory check now includes size of page tables. This can be important e.g. on ORACLE systems with a lot of memory
• windows_agent: Now buffers output before writing it to the socket Results in less tcp packages per call
• smart.stats: rewrote check. Please reinventorize. Error counters are now snapshotted during inventory.
• smart.temp: add WATO configuration
• windows_agent: check_mk.ini: new option "port" - specifies agent port
• winperf_processor: introduce averaging, support predictive levels
• cpu_util.include: fixed bug when params are set to None
• predictive levels: fixed bug when existing predictive levels get new options
• windows_plugin mssql.vbs: No longer queries stopped mssql instances
• cisco_hsrp: fixed problem when HSRP groups had same ip address
• winperf_if: hell has frozen over: a new check for network adapters on Windows
• windows agent: new config section plugins, now able to set timeouts for specific plugins new global config option: timeout_plugins_total
• lnx_if in Linux agent: force deterministical order of network devices
• Linux agent: remove obsolete old and sections
• logwatch, logwatch.ec: detect error in agent configuration
• Linux agent: cups_queues: do not monitor non-local queues (thanks to Olaf Morgenstern)
• AIX agent: call lparstat with argument 1 1, this give more accurate data
• Check_MK check: enable extended performance data per default now
• viprinet checks: New checks for firmware version/update, memory usage, power supply status, router mode, serialnumber and temperature sensors
• uptime, snmp_uptime, esx_vsphere_counters.uptime: allow to set lower and upper levels
• winperf_processor: Now displays (and scales) to number of cpus in pnpgraph
• mk_postgres plugin: replace select * with list of explicit columns (fix for PG 9.1)
• lnx_if: show MAC address for interfaces (needs also agent update)
• winperf_tcp_conn: New check. Displays number of established tcpv4 connections in windows Uses WATO Rule "TCP connection stats (Windows)"
• windows_agent: fixed timeouts for powershell scripts in local/plugins
• logwatch: Agent can now use logwatch.d/ to split config to multipe files
• logwatch: Agent can now rewrite Messages
• apache_status: New rule: set levels for number of remaining open slots
• mrpe: handle long plugin output correctly, including performance data
• Notifications:
• notify.py: unique spoolfiles name no longer created with uuid
• Warn user if only_services does never match
• Livestatus:
• Table statehist: Improved detection of vanished hosts and services. Now able to detect and remove nonsense check plugin output
• FIX: able to handle equal comment_id between host and service
• livestatus.log: show utf-8 decoding problems only with debug logging >=2
• livestatus: fixed incorrect output formatting of comments_with_info column
• BI:
• Integrated availability computing, including nifty time warp feature
• WATO:
• Configuration of datasource programs via dedicated rules
• New editor for Business Intelligence rules
• Rule Editor: Now able to show infeffective rules
• Valuespec: CascadingDropdown now able to process choice values from functions
• cisco_qos parameters now configurable via WATO
• Removed global option logwatch_forward_to_ec, moved this to the logwatch_ec ruleset. With this option the forwarding can now be enabled for each logfile on a host
• Configuration of an alternativ host check command
• Inventory: Display link symbol for ps ruleset
• New rule for notification_options of hosts and services
• FIX: Rulesets: correct display of rules within subfolders
• Remove Notification Command user settings, please use flexible notifications instead

What is new in version 1.2.2:

• Core:
• Added $HOSTURL$ and $SERVICEURL$ to notification macros which contain an URL to the host/service details views with /check_mk/... as base.
• Checks & Agents:
• FIX: blade_bx_load: remove invalid WATO group
• FIX: lnx_bonding: handle also 802.3ad type bonds
• Notifications:
• FIX: Removing GRAPH_CODE in html mails when not available
• Using plugin argument 1 for path to pnp4nagios index php to render graphs
• Little speedup of check_mk --notify
• Multisite:
• FIX: Fixed umlaut handling in reloaded snapins
• WATO:
• FIX: Fix several cases where WATO rule analyser did not hilite all matching rules
• Added tcp port parameter to SSL certificate check (Thanks to Marcel Schulte)
• Event Console:
• FIX: Syslog server is now able to parse RFC 5424 syslog messages

What is new in version 1.2.2 Beta 6:

• Checks & Agents:
• FIX: postfix_mailq: fix labels in WATO rule, set correct default levels
• WATO:
• FIX: syntax error in check_parameters

What is new in version 1.2.2 Beta 3:

• Checks & Agents:
• FIX: Fixed date parsing code ignoring the seconds value in several checks (ad_replication, cups_queues, heartbeat_crm, mssql_backup, smbios_sel)
• FIX: Fixed pnp template for apc_symmetra check when using multiple rrds
• Multisite:
• FIX: Removed uuid module dependency to be compatible to python < 2.5
• FIX: remove Javascript debug popup from multi-string input fields
• FIX: list of strings (e.g. host list in rule editor) didn't work anymore

Requirements:

• Python
• Nagios