October Release – v.54

A lot of development that’s been done is not yet visible, either as it’s dependent on the new UI, or simply spanning several releases as it involves a large project.  In the near term, we can all look forward to greatly improved Netflow; trend forecasting; integrated AWS services monitoring; and, of course, the completion of the new UI.
But for this release, there are still some goodies:

Improvements

  • Datasource Improvements:
    • Increased the limit for embedded groovy scripts in active discovery or script collection to 64,000 characters.
    • Instance filtering now applies to JDBC AD methods for discovery type of “database list” and “instance list”
    • Added a “delete obsolete instances immediately” flag for datasource Active Discovery, useful for when the history of prior instances is unrelated to new instances of the same name.
    • XML Document Xpath interpretation for datapoints has been enhanced to support all Xpath core functions (sum, count, etc)
    • JsonPath interpretation has been enhanced to support sum, average, min, and max functions for a JsonPath. Regex filtering support was also added.
    • Added pdh.user and pdh.pass properties. If these are set, they will be used for Perfmon data collection authentication. If they are not set, wmi.user/wmi.pass will be used. There are rare cases where it is necessary to specify different credentials for each protocol (typically where monitoring outside a domain.)
  • Collector Improvements:
    • Improvements in the performance of ping collector – the first to use a much more scalable system, that you’ll be seeing in other collection mechanisms soon.
    • Added the ability of collectors to detect and download updated scripts, when the script associated with a datasource has changed content, but the name stays the same. (Previously, the script would only be re-downloaded if it was called with a new name.)
    • Added the ability for host properties to be passed to external scripts used for collector based alerting.
    • IPv6 is now enabled by default
  • Reduced the size of the LogicMonitor logo on reports, to avoid brand confusion for those MSPs sending these reports to their customers.
  • Added compound Windows Event filtering

Bugs resolved

  • Collector issues:
    • Uninstalling a Linux Collector service would not correctly terminate the running collector.
    • Removed a race condition between watchdog and the collector wrapper, that could result in a failed restart
    • Improved the ability of scripts called by external alerting to deal with white space in parameters and path
    • Collector SDTs that spanned two days would sometimes not suppress alerts correctly.
    • Alert clears were sent for collector down states, even though the collector had been placed in SDT.
    • When SDTing a collector for which the down state had already been acknowledged, the SDT notifications were incorrectly sent.
  • Custom HTTP delivery methods (Webhooks) could deliver illegal JSON if the content contained tokens that were replaced with text with newlines, tabs, etc. These characters will be replaced with spaces.
  • Sitemonitor did not deal with 204 HTTP responses correctly.
  • The filter “cleared(last 7days)” was not working correctly on alert/service page in the new UI
  • A netscan using a script method incorrectly added the host using the SNMP discovered name, instead of the name specified in the script.
  • Corrected a bug where script based network scans that specified a group to add hosts to, would fail to add hosts to the specified group if the scan was run by a Windows collector
  • Corrected a few Role Based Access control bugs: Users with View only privilege on hosts were incorrectly able to modify cluster alerts and event sources on a group; the Alert Report was not respecting host limitations.
  • The Alert Threshold report would erroneously show a specific instance when thresholds were set on group or global levels, instead of showing they were applicable to all instances.
  • Corrected an issue that caused discovery filters for JDBC AD to fail if the filter referenced a column ‘name’.
  • Prevented Null/Empty responses for OS system properties from replacing prior discovered system properties in the Auto Properties discovery system
  • Corrected an issue that resulted in SLA reports of duration greater than one hour reporting on an extra minute.

In this Article: