Welcome to docs.opsview.com

This is an old revision of the document!


This page lists the major changes between releases

Opsview 4.6.0


  • Security Wallet - all passwords are now hidden from the user interface and encrypted in the database. Connections to databases can optionally be encrypted. It is not possible to retrieve passwords in clear text from the REST API. Opsview slaves will also have unique encryption keys that are not the same as the master's. More information in our documentation
  • Attributes can now be encrypted for user specified data
  • New platforms supported: RHEL 7, CentOS 7 and Ubuntu 14.04
  • Support for sFlow devices in addition to NetFlow


  • Monitoring Plugins are now used in preference to Nagios Plugins
  • Perl modules upgraded to Aug 2014
  • Improved error message on SNMPv3 test connections where privacy password or privacy protocol is incorrect
  • Added an include_ms=1 for /rest/config/host to identify if a host is a master or used as a monitoring server
  • Dashboard events timeline dashlet uses a new dynamic refresh counter
  • REST API for hosts can optionally include SNMP interfaces


  • As part of this upgrade, we remove any NetFlow source that does not have a valid IPv4 address
  • Due to changes to support password encryption, sensitive data is no longer returned via the REST API, such as SNMP credentials and encrypted host attributes. This may affect clients that expect the data to be present
  • SNMP traps will now get recorded with the community string hidden and set to 5 asterisks
  • Resolve service checks page will be removed in this release, but a new enhanced version will be available in the next release
  • On upgrade from previous versions of Opsview, attributes will not be automatically converted to be encrypted. See documentation for more information
  • REST API validation rules have been tightened for the labels of SNMP polling service checks
  • Apache is now restarted on all platforms as part of Opsview Web install - make sure that Apache is correctly configured before upgrading
  • check_snmp_interfaces has a different structure for state data, so you may get an initial alert for differences
  • check_http_authentication now checks SSL certificates by default - disable with –no-certificate-checking
  • CentOS 5 and Ubuntu 10.04 Lucid are no longer supported for Opsview master
  • IE8 and below will get a warning message on the login page as these versions of IE are no longer supported
  • Safari and IE8 are deprecated browsers


  • Restored RRDGRAPHS as an access point for Public roles
  • Remove ndolog files if a rename fails, used during debugging
  • Fixed autodiscovery issue when running VMware scan from a slave, including capturing stderr for debugging
  • Fixed empty list of host templates when editing service checks
  • Fixed Dashboard issue where cannot add a shared dashboard with UTF-8 characters
  • Fixed Dashboard where discrepency between Keyword List view and configuration list (OP-268)
  • Fixed Dashboard Keyword List and Top Service Issues sorting of status, so critical > warning > unknown > ok
  • Fixed Dashboard Top Services by Metric so that always lists the highest metric values, displays values as floating points and subsorts by hostname
  • Fixed Dashboard events views where filtering of host was ignoring service events and vice-versa
  • Fixed Dashboard stacked graphs where a metric of 0 was causing the entire stack to be empty
  • Fixed Dashboard issue showing BSM Dashlets with N/A for availability when it should be 0% (OP-1395)
  • Fixed Dashboard process map unnecessarily loading background image (OP-598)
  • Fixed smbclient dependencies for check_disk_smb
  • Fixed query host issue where interfaces with trailing white space were not getting matched with existing selected interfaces (OP-2531)
  • Blocked sorting of unsortable columns in events view (OP-2979)
  • Fixed issue with BSM configuration where Update button would not be enabled (OP-3107)
  • Fixed issue where BSM services/components not listing beyond first 50 records (OP-3478)
  • Fixed Nagios allowing acknowledgements when host/service already acknowledged
  • Fixed Nagios issue where notifications sent out even though downtime already has been set. Only occurs just after a reload where lots of results are sent via the external command pipe
  • Fixed check_http_authentication where warning and critical thresholds were not being respected correctly