GD Collector 23.0 will be available on February 8, 2017. This version will be the foundation for our next mandatory version, and includes everything in MGD 22.005 as well as EA 22.050, EA 22.084, EA 22.110, EA 22.142, EA 22.180, EA 22.181, EA 22.227, EA 22.228, EA 22.253, and EA 22.276, plus one bug fix since EA 22.276:
- Services configured to make HTTP HEAD requests weren’t always working correctly with EA 22.276.
In comparison to MGD 22.005, highlights of GD 23.0 include:
- Support for LogicMonitor’s Internal Service Checks.
- Collector Caching – This enables the Collector to store data in your environment during periods where your collector is unable to deliver data to your LogicMonitor account. By default, Collector caching is enabled and configured to cache up to 30 minutes of data. See documentation for more information.
- Instance Level Properties – Allow you to set properties at the datasource instance level, or have them discovered via Active Discovery.
- Collector Sizes – Different sizes correspond to different JVM max memory sizes, and make it easier to ensure that your Collector is configured to handle its intended monitoring load.
- Script Output Logging – For collector.script log levels of debug and trace, we now log script data collection and active discovery stdout and stderr in the agent/logs directory.
Discover AWS Tags for EC2 Instances – We’ve added Collector support to discover and update AWS tag values for EC2 instances added into monitoring via Netscan. Consistent with how properties are typically updated, our Auto Properties subsystem will automatically run and update these properties on a periodic bases.
Upgraded external libraries:
- vijava to yavijava 6.0
- Windows NTPQ.exe
- AWS SDK to version 1.11.56
- Jackson jars to version 2.6.6
- Maria JDBC Driver to version 1.4.5
- In certain cases, HTTP response times were sometimes reported inaccurately, showing a faster response time than actually existed. This was because HttpClients & cookies were shared across asynchronous HTTP collection tasks. Cookies will now be stored per HTTP task, which resolves the issue.
- In some cases, increasing log levels such that a large number of messages were logged was causing the Collector to restart frequently. We’ve optimized how the Collector handles such volumes of log messages to correct for this.
- Webpage data collection wasn’t working for sites with certificates using SHA256withDSA as the signature algorithm. This was due to a known issue with JDK 8. We’ve upgraded the Collector to JDK 8u102, which includes a fix for the issue.
- ESX data collection wasn’t working for servers where OpenSSL was upgraded to 1.0.1t. This was due to a cipher suite used by JDK 8 and disabled in OpenSSL 1.0.1t. We’ve corrected the issue by adding an option to disable the relevant cipher suites on the Collector, only for the ESX collection mechanism. This option is enabled by default.
- We’ve removed the Metaspace memory limit and improved how the ##WILDVALUE## syntax is handled in Groovy scripts, decreasing the amount of memory needed.
- We’ve disabled the creation of mdmp files that previously occurred when the Collector crashed because of an out of memory exception.
- Netflow improvements for large amounts of Netflow traffic.
- JMX, WMI, HTTP, and Perfmon data collection improvements, resulting in an increase in collector capacity for these protocols.
Note that this Collector no longer supports CentOS 5.
For more detailed information on what’s included, see individual EA release notes (links above).