Page History

...

• opEvents' timezone handling was greatly extended.
  If you set the config option omkd_display_timezone to your desired timezone, then all times in the opEvents GUI will be displayed in that timezone and including the timezone offset.
  You can use any timezone definition from the ISO8601 standard and the Olson database, plus "local" (meaning the timezone configured on the server).
  If this option is not set, the times will be shown in the "local" timezone but without zone offset. If explicitely set to "local", the offset suffix is shown.
• opEvents now supports times with timezone specifications in the advanced search dialog (but  only in numeric offset format at this time, e.g. "+0500").
• opEvents now records the original event date property (if any) separately from the underlying raw epoch time, and timed records for all of an event's processing stages are recorded as well.
  All of these are shown on the event details page.
• Escalation policy actions are now shown more prominently in the action log display.
• The handling of special characters in policy action substitutions was improved, and the example EventActions policy file updated.
  Please note that the EventActions file shipped with version 2.0.3 is insufficiently robust and should be replaced with the new version at your earliest convenience.
• The log handling was improved. Log reopening works more reliably, and opEvents daemon logs are now prefixed with the component role and process id.
• Improved robustness for the rest-style API for remote event management and the example client application.
• New mechanism for displaying a dynamic service priority text (by event tagging with servicePriority).
• Improved robustness and efficiency for MongoDB operations.
• opnode_admin is now able to clean up inconsistent semi-existent nodes, including events for that node.
• opEvents can now optionally ignore events for unknown nodes.
  This is configured using the configuration setting opevents_auto_create_nodes, which defaults to true if not present. If true, node records are automatically created if necessary.
  If set to false, no nodes are automatically created and events for unknown nodes are completely ignored.
• The GUI authentication expiry can now be adjusted with the configuration setting auth_expiry_seconds.
• Various opEvents GUI pages were adjusted for improved usability and better access to events' context and details.
  For example, the event context for stateful events now includes links to any related/opposite event.
• opEvents now performs policy actions email and script asynchronously and in separate processes.
  This speeds up event handling substantially because the main event reader process does no longer have to wait until the programs that your action policy triggers actually do finish.
  (For example, diagnostic programs like traceroute can easily take 30 seconds to complete.)
  If an event has actions pending processing or completion, a notice info bar is shown on the event context page.
  The new config option opeventsd_max_processes lets you set a limit for parallel worker processes; if that limit is exceeded, further action processing is queued and performed later.
• Storm control and event correlation capabilities were improved.
  Both programmable event suppression and event correlation polices now support the option to automatically acknowledge the suppressed/triggering events. Furthermore they also allow the optional delaying of event policy actions for a configurable extra delayedaction period.
• opEvents now supports high-precision timestamps better, and displays both human-friendly and raw time stamps on the event details page.
• Stateless events sourced from NMIS' event log (or slave poller log) are handled more consistently and robustly, and only events with both state and stateful properties are interpreted as stateful.

...

• opEvents 2.0.0 requires NMIS 8.5.10G or newer for full interoperability. Please check the Product Compatibility page for details.
• Improved license management user interface
  It's now much easier to delete, restore or selectively import your licenses from your Opmantek.com account into opEvents,
  and reminders about any license conditions being exceeded are presented in a more useful fashion.
• There is a totally new rest-style API for remote event management, complete with an example client and its source.
• Configurable purging of old data from the opEvents database is now supported.
• opEvents provides a new facility for summary reports (created both on demand and periodically), complete with automatic email of reports as XLSX spreadsheet and online display.
• Comments with attribution and timestamps can now be added to nodes and events.
  Anybody who can view an event can also add a comment for it, but only an administrator can delete comments. (Node comments are all admin-only.)
  Comments are shown with the event or node in question, and are tagged with the creation time and originating user. Any URLs in comments are presented as clickable links.
  The older facility for importing and editing notes for nodes still remains.
• More user-friendly new landing and overview page at "http://<yourserver>/omk".
• Improved interactivity due to better database connection caching
  opEvents now keeps its connections to MongoDB open and reuses them as much as possible, which improves interactivity especially if you use a remote MongoDB server.
• New Node Overview screen
  There is a new node-centric dashboard or overview screen which shows events and event types for a node over time. Links between this node overview and the node context allow easy navigation.
• Various GUI improvements and  refinements, e.g. more informational window/tab titles.
• Improved NMIS importing now also covers opHA-slave poller nodes, and access to interfaces' ifAlias property
  Importing or refreshing  nodes from NMIS now handles nodes on remote slave poller instances if opHA is active on the opEvents server. Event action rules can now access an interface's Description or ifAlias property.
• Events can now include links to other "authoritative sources", e.g. external applications like helpdesk systems or the like.
  (See the documentation about authority and location properties in the list of Normalised Event Properties for details.)
• Node editing actions are now logged with timestamp and originating user in logs/audit.log
• Improved access control, better NMIS authentication integration
  opEvents now fully enforces access control based on a user's group memberships: only those nodes and events are visible, where the nodes are members of groups that this user is authorized to see.
  The installer now also offers to merge NMIS and Opmantek users.dat password files.
• Better logging and log-rotation support
  opEvents now logs to log/opEvents.log and the log format and content was revised to make the logged information more useful. Logs are reopened when the opeventsd receives a SIGHUP signal.
• opEventsd now restarts automatically when any relevant configuration files change.
  opEventsd now can be instructed to also restart periodically, using the new opeventsd_max_cycles configuration directive (= restart after so many opeventsd_update_rate intervals).

...

• opeventsd now resurrects its worker slave poller on demand
• a new generic parser engine for local customisation was added
• opeventsd can now create events from the command-line or from a single JSON file
• opevents now supports any number of logfiles per log parser type
• the ip address cache now works properly and more efficiently
• opeventsd now displays all supported command-line arguments when run with -h or -?
• opevents now properly handles truncated (not rotated) logs