Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

This is a brief overview of the major changes between opCharts releases.

...

  • Metrics and Health Component now works for RBAC users.
  • You can now create custom charts on the master Primary with poller datasets
  • You can now search by regex in the node edit screen

...

Released 25th June 2020 as General Release

(warning) This release requires updated licenses, please contact Opmantek Support to organise new licenses

Highlights

All OMK configuration are now JSON, the processes to convert from NMIS configuration files to JSON is performed by the installer, click here for more details about JSON Configuration..

...

  • The Metrics and Health component has been extended to group by any node configuration property.
    • This is set in opCommon.nmis opcharts_metric_summarise_by, you need to define an array of hashes, with key and name
    •     {
            'key' => 'group',
            'name' => 'Group'
          },

    • If more than one selector is defined on the top left of the Metrics and Health component a dropdown is shown to change which confiuration item the nodes are grouped by.
    • Metrics and Health gauges are now rendered in one row which are scroll-able to the right.
    • Metrics and Health links now correctly affect the nodes filter when clicked.
    • Metrics and Health is now searchable on the currently selected column.
  • New Event History component which utilizes the NMIS9 events DB
    • For this to work in opCharts you need to set keep_event_history => 'true' in NMIS9 Config.nmis
    • Consideration should be given to the volume of data on the pollers and master Primary and how much event history will be kept, by default NMIS9 will keep 7 days of event history which is configured in Config.nmis/purge_event_after (7*86400), if having problems viewing event history, reduce the number of days and purge the data.
    • This component replaces the old tail log event viewer but if you want to use the old component change opcharts_db_event_history to 0 in opCommon.nmis
    • Currently websockets are not enabled, this component is set to refresh period
  • Revised Events screen to utilize the new Event History component
  • Fixed issue where adding a group to a map would cause the node selection modal to appear
  • Fixed issue where the nodes interfaces panel which shows interfaces with oper status down would fail to render
  • Fixed issue where the nodes interface component would incorrectly show certain interfaces
  • Fixed issue where the node selector on the nodes screen will fail to render if a group name is numeric
  • Fixed issue where the TopN table text would not correctly space the table contents
  • Fixed issue where a node info panel would cause a 500 if the node does not exist

...

Released August 22th, 2019
opCharts 4.0.7 requires NMIS 9.0.6

  • Fix TopN query error in master Primary servers with lots of data from the poller servers. The error was a MongoDB::DatabaseError: operation exceeded time limit.
  • New Admin function to set many user's default dashboard:
    • ./opcharts-cli.pl act=import-user-dashboard name=dashboard-name users='user1, user2, ...'

  • Fixed changing the default dashboard for a user in the GUI. 
  • NMIS button works for non local servers now: Redirects to the NMIS node poller view. 
  • GUI improvements:
    • Fixed size in NMIS button (Nodes View) and set visible as default.
    • Improve view for long group names in Metrics and Health view, not overflowing.
    • Improve the view for long names in TopN table.
  • Support for watching monitored services graphs from the poller servers in the masterPrimary
  • Fixed bug that prevented hiding ssh button in node info panel where it was disabled using opcharts_gui_display_node_ssh.
  • Fixed bug that was failing to check permissions for a subnet.

opCharts 4.0.6 Beta

Released July 19th, 2019
opCharts 4.0.6 requires NMIS 9.0.4

  • Metrics and Health component has been rewritten to support Master Primary and Poller, this is done over the whole active nodes collection on the current server.
    • When you have nodes from different pollers Gauges are used to shows the current groups Reachability
    • If you want to change which metric the gauge can display copy omk/lib/json/opCharts/components.d/nmis_network_health.json into omk/conf/components.d
      • change "gaugeDataSource" to one of the desired values
        • avg_available or avg_16_available
        • avg_health or avg_16_health
        • avg_reachable or avg_16_reachable
        • metric or metric_16
        • status
      • Change "gaugeTitle" to a desired title
    • In poller mode where only the current polled nodes are on the server Metrics and Health will show the current NMIS Network Health chart
  • Added "websocket_proxy_port" to opCommon, this can be used to override which port the web client should connect its web-socket to.
  • Node searching has been rewritten, now much faster.
    • You can search over the nodes
      • Name
      • Host
      • Group
  • Fixed issue where the Node list could not be exported to csv. Use opcharts_node_list_exporttocsv to configure which fields to export
  • Fixed issue where some pages would not resize correctly when hiding the sidebar
  • Fixed node panels KPI colouring by level.
  • View node on local NMIS now correctly uses the nmis9 URL defined in opCommon
  • Added View Nodes on poller, more can be found Here

...

  • Fixed issue where a node name is numeric fails to display in a geo map
  • When you are shipping poller logs to the master Primary these can be displayed for the node my changing the config option 'opcharts_event_log_type' => 'poller_event_log' 
    • These logs are expected to be in the default NMIS9 directory for logs <nmis_logs>/poller_event.log 
    • NOTE: refer to Config.nmis for your setting of <nmis_logs>, it defaults to /usr/local/nmis9/logs/
  • If you wish to hide the event log widget for nodes you can set 'opcharts_gui_node_eventlog_display' => 'false'
  • Fixed issue where 'auth_default_groups' => 'all' would not allow access to 'all' group
    • Set auth_default_groups & auth_default_privilege in opCommon to use this feature.
  • Monitored services: This view will show monitored services from the poller nodes. 
  • Bug fix to show IP address and mask on inventory and interface list. 
  • Node panels now correctly show Health Kpi and Response Time
  • In an MSP environment and wish to hide popup notification errors to users bar users with administrator privileged use 'omk_gui_show_user_errors' => 'false'

...

If you need to sort by node health in the nodes list while using DB caching you need to set "nmisx_db_cache_node_health_sorting" to true, this setting was introduced in 3.3.8 and is false by default.

(warning) When using opCharts as a Master Primary and with "nmisx_db_cache_enable" set to true, data in the nodes list, which is derived from a pollers node summary8h or summary16, cannot be displayed

...

  • Node info panel (found on most node pages) now allows for more complex formatting, eg. displaying a unix timestamp as human readable, see documentation for more
  • Delegated Authentication
  • Dataset charts now show Min/Max/Avg
  • Custom JS
  • Link to opFlowSP Agents/Interface from Node Interface view

Fixes

  • Master Primary redirect to poller with RBAC user now keeps context_list so user can see URL

  • Users/Roles with "." (period) in the name can now be edited/deleted
  • NMIS Group search function now functional

...

  • Authorisation fixes for dashboards/business services

  • Group links in maps fixed

  • Charts once again honour line type setting
  • Chart creator no longer displays first dataset twice in chart
  • Node list table mode now links to pollers properly

  • Node list sorting error fixed

  • KPI coloring now more consistent between table and panel

  • Dashboards with external pie’s and charts now load properly

  • Table/page size node list page size should now be consistent

  • Attempting to access poller node detail on master Primary now automatically redirects to poller

  • SSO cookies with domain that is too specific (sometimes)

...

  • Show only known Monitored Services when not in master Primary mode
  • Fixed issue with deleting entities on map
  • Network maps no longer have bottom 20px of background clipped
  • Loading and no data found notifications for charts fixed
  • Consistently show % in TopN tables

...

  • Maps can now hold nodes from pollers
  • TopN only considers data from the last 15 minutes (so old data will be ignored), time period is configurable
  • Maps holding nodes a user is not authorised for appear grey (but are still on map)
  • More uniform support for masterPrimary/poller nodes 

Fixes

  • Users no longer see nodes in node search that they cannot access
  • Better at handling blank nodes in NMIS
  • TopN report only runs once every 5 minutes (configurable) no matter how many times it is called (force option available)
  • Dashboard viewing now verifies a user has rights before displaying anything
  • Pressing new button on auto-generated charts fixed
  • Verified whole system to ensure nodes only visible to users who have rights to see them

...