EA Collector 22.2227 will be available early next week, and includes the following:
- In EA 22.180 we added an option to disable certain cipher suites on the Collector for ESX data collection, to account for a conflict between JDK 8 and an OpenSSL security patch. This option defaulted to false – we’ve defaulted it to true in EA 22.227.
- Updated the Collector’s Windows NTPQ.exe. The previous version had a bug that, in certain cases, resulted in inaccurate values for devices reporting no data.
- We’ve improved how the Collector’s Auto Properties system detects SNMP version. We previously checked SNMP versions serially, which could result in longer than usual Auto Properties executions. EA 22.2227 will now check SNMP versions in parallel, and set the SNMP version to the highest responsive version.
- EA Collector 22.084 introduced a feature to log script output to file. We’ve made the naming of these files more consistent, specifically ‘datacollecting-HostName-DSI.stdout’ for script data collection & ‘discover-HostName-DATASOURCENAME.stdout’ for script Active Discovery.
- JMX data collection improvements
- We’ve disabled the creation of mdmp files that previously occurred when the Collector crashed because of an out of memory exception.
- The JSON post processing method previously only worked for JSON objects, not JSON arrays. We’ve improved this post processing method to work for JSON arrays.
- When AWS tags are removed from resources, any associated properties will now be removed in LogicMonitor.
- Updated the Collector’s AWS SDK to version 1.11.56
- Updated the Collector’s Jackson jars to version 2.6.6
- Previously, Expect.open did not have a default timeout configured. This resulted in failed scripts tasks in a few edge cases. With EA 22.227 we are setting a default timeout of 10 seconds to avoid the issue.
- You can now use ILPs generated via Script Active Discovery in AD Filters. Just reference the ILP via name, e.g. auto.instanceSerialNo. Note that SNMP and WMI generated ILPs can also be used in AD Filters with the same syntax.
- Older Collectors without ‘snmp.timeout’ and ‘snmp.threadpool.size’ options in the agent.conf file weren’t upgrading correctly to EA 22.142. We’ve fixed this issue in EA 22.227.
- SNMP Active Discovery filtering wasn’t working correctly when the OIDs specified in the filter timed out. This resulted in instances passing filters where they shouldn’t have.
- The md5RSA signature algorithm was not enabled by default for Internal Service Checks, causing SSL error in certain cases. This algorithm is enabled by default in Collector version 22.2227.
- The Collector’s sbshutdown process was previously configured to shutdown all processes named “java.exe”,”sbwinproxy.exe”, & “wrapper.exe”. If another java.exe process was running on the server, it was shutdown by sbshutdown. In EA 22.227 we’ve added logic to verify that the processes being stopped are the Collector’s processes.
- Previously, webpage Collector response properties could only be referenced in upper case syntax. This was only an issue for groovy complex datapoints that referenced these properties in lower case syntax (e.g. response “body” instead of “BODY”), and only applies to the webpage collector. EA 22.227 makes these properties case insensitive.
- For sites with more than one certificate, we previously reported the soonest certificate expiration. In certain cases, this resulted in unnecessary SSL errors. The Collector will now report the latest certificate expiration for one site with more than one certificate.
- In cases where JDBC queries returned no matching records, the Collector would not report response time and status. We’ve changed this logic such that status & response time are reported even when no records are returned.
- Script EventSources using embedded Groovy weren’t interpreting properties passed between ## correctly.
- Scripts using the Groovy Expect API and handling outputs larger than 65536 bytes were resulting in No Data in a few rare cases. With EA 22.227 we’ve optimized how we handle such outputs to avoid the issue moving forwards.
- The Collector wasn’t getting data from MXBean objects correctly in some cases.