Each of our Resource Manager releases focuses on a theme. For this release, the chosen theme delivers on ZenHub Improvements. ZenHub is the internal component that governs data flowing to and from the collectors. Improving ZenHub will bring meaningful improvements to our customers’ ability to deliver high levels of service to their organizations.
In prior releases, ZenHub delegated much of its work to subprocesses running in the ZenHub service’s container. Customers could scale ZenHub processing power by increasing the size of its container.
With RM 6.3, we have implemented horizontal scaling, improving overall ZenHub performance by increasing the number of containers that can handle ZenHub processing tasks. This enhancement will leverage Control Center to provide horizontal scaling.
This enhancement will have these visible changes:
- Control Center will replace the single ZenHub service with two services: ZenHub and ZenHubWorker
- The ZenHubWorker service can be configured to have multiple instances
- The ZenHub service will be restricted to a single instance
The effect of these changes means that users can distribute ZenHub processing across all hosts in the delegate pool.
Some organizations have previously built a “superhub,” using a separate ZenHub-only delegate pool with dedicated resources. With this enhancement, the superhub approach should no longer be necessary in most cases.
Other notable Resource Manager changes include:
- Customers can investigate device modeling issues with a new user interface option for debug modeling. This function eases resolution of device loading issues, improves ZenPack development with faster access to detailed debugging logs, and improves security by eliminating a common requirement for direct container access.
- When device and component graphs are viewed over long time periods, several raw data values are aggregated to create one data point on the graph. With this release, we provide the ability to control aggregation at graph view time. Customers can choose mean, low, high and 95th percentile aggregation.
- Reportmail now supports delivery of graph reports.
- Support for the upcoming Service Impact 5.4 release. Service Impact 5.4 implements selective synchronization to improve performance and reliability by dramatically reducing the size of the Impact graph database. Initial customer-environment test results are very favorable.
~80 customer-requested usability, reliability, serviceability and performance changes.
Note: The example screenshots above are from a Zenoss Cloud tenant. Upgrades to ZSD and Control Center are deployed in Zenoss Cloud as well.
Control Center is also updated:
- Automatic remediation of deadlocked Zope services improves Zenoss availability. Added support to restart Zope services if a health check times out three times in a row.
- Corrected a medium-severity security issue by requiring Zookeeper authentication for delegates (https://www.cvedetails.com/cve/CVE-2018-8012/).
Beginning with this release, we publish ZSD documentation first to https://zsd-help.zenoss.com. This is the same documentation system we are using for Zenoss Cloud. Documentation for prior versions of ZSD will be available on the ZSD help site. A key benefit of the new documentation system is that we will be closely integrating traditional documentation and Knowledge Base articles for better information delivery.
We recommend that all existing customers upgrade to these releases. Customers upgrading from version 5.x can significantly reduce upgrade time by following steps in this Knowledge Base article. For further information, see the Release Notes on the Zenoss documentation pages.