1.40. Release Notes 0.967

1.40.1. Important changes in this release

1.40.2. New features and improvements in UI

  • view “all” is now optional. Default view builder Python script creates this view, but users with large networks where this view produces huge and unusable map can create their own view builder script without this view.
  • Devices are always “pinned” in all maps, however user can adjust their positions manually. This eliminates device “floating” in maps.
  • Limited support for making map layout specific to authenticated user has been added: This only works when NetSpyGlass is deployed behind authenticated http proxy that injects HTTP header with user name. Configuration parameter “ui.authentication.reverseProxyAuthenticationHeader” can be used to set up regex pattern to match this header line and extract user name. See document doc/authentication.md for more details and examples.

1.40.3. New features and improvements in the Monitor

  • Monitor spreads SNMP queries out to reduce CPU load on the device. Instead of sending all queries at once, they are spread evenly over the polling interval period.
  • Default polling interval has changed and is now 1 min
  • default RRD configuration is now as follows:
    • original data points, keep for 24 hours
    • 5 min averages, keep for 7 days
    • 1 hour averages, keep for 90 days
    • 1 day averages, keep for 1 year
  • improvement in the default variable selector Python script: it won’t monitor traffic, errors, discards on Juniper subinterfaces that are aggregation ports. Traffic, errors and discards are monitored on the corresponding parent physical interfaces.
  • Improved error reporting for situations when the program can not load Python hook script or can not load corresponding class.
  • HBase row key generator algorithm has changed in this version to make row keys independnt from the polling interval. This only affects new data stored in HBase with NetSpyGlass v0.967 and later. The application can retrieve data stored with old row keys as well but will not use them to store new data. User can change polling interval at any time (application restart is required) and application will be able to retrieve both old and new data from HBase.
  • Monitor adds “synthetic” monitoring variable for each device that tracks percentage of SNMP queries that time out. This variable has name snmpTimeoutsPercentage, it appears in Device Details panel and Graphing Workbench.

1.40.4. Notable Bugs fixed in this release

  • Juniper interface “em” used to be considered “out of band management” by NetSpyGlass and ignored. Starting with v0.967, this is no longer the case, these interfaces will appear on maps and can be monitored.
  • Monitor used to return latest observation value for monitoring variable even if it was collected some time in the past. For example this could happen if interface went down - the monitor would keep reporting the latest value of the traffic level, errors and discards even though no new data was being collected. Now the monitor returns latest observation only if it is no more than two polling cycles in the past.