OpenStack is a massively scalable, feature rich open source cloud computing platform for public and private clouds. Corporations, service providers, VARs, SMBs, and global data centers all use OpenStack to deploy large scale private and public clouds.
Zenoss has been providing monitoring support from a tenant perspective for OpenStack clouds for the last several years. However, over the last several months, the Zenoss engineering team has been hard at work enhancing and extending our existing OpenStack monitoring capabilities to now include a new provider-oriented approach to OpenStack performance and availability monitoring. (See our announcement Zenoss Increases Return on Investment for OpenStack-Enabled Clouds for more details.)
OpenStack Monitoring Challenges
OpenStack clouds provide a lot of great functionality. However, as many Zenoss customers who are already using OpenStack know, getting an OpenStack cloud up and running can be challenging. With OpenStack, everything is very piecemeal — distributions, projects, components, etc. It can sometimes feel like you need an army of developers to get an OpenStack cloud rolled out. And even after you do get an OpenStack cloud rolled out, the work is far from over — you are still going to need a pretty experienced team in order to support an OpenStack cloud in a production environment.
Your OpenStack cloud is mission-critical because it is the basis for computing services you are providing to your internal or external customers. So, just like traditional enterprise computing infrastructure, you need to monitor your OpenStack infrastructure in order to assure the health of those services. However, the state of OpenStack monitoring technologies is, like OpenStack itself, still maturing.
If you have an OpenStack cloud you need to monitor, Zenoss can help. You can use Zenoss to monitor your OpenStack cloud, as well as all of the rest of the components in your enterprise IT infrastructure — whether they are physical, virtual, or cloud — using a single, unified view that helps you improve your service quality and reduce your IT operational costs.
New OpenStack Infrastructure ZenPack
Zenoss provides a very unified, holistic approach to IT infrastructure monitoring, including OpenStack cloud monitoring. This includes providing both tenant-oriented, as well as provider-oriented, OpenStack cloud performance and availability monitoring.
Our existing OpenStack Cloud Monitor ZenPack focuses on providing performance and availability monitoring of OpenStack “flavors”, “images”, and “instances” from the perspective of an OpenStack consumer, or tenant. However, Zenoss has recently released a new OpenStack ZenPack, the OpenStack Infrastructure ZenPack, which focuses on performance and availability monitoring from the perspective of the OpenStack provider, or administrator of the OpenStack cloud, and includes new provider-oriented metrics. This new ZenPack was specifically developed to take advantage of some of the new features available in OpenStack Icehouse, the ninth release of OpenStack, including new capabilities provided in the Compute (Nova) and Metering (Ceilometer) projects.
In the new OpenStack Infrastructure ZenPack, we’ve focused on leveraging and integrating with the Nova, Ceilometer and Keystone projects to provide a single view into the overall OpenStack environment. The current release of the ZenPack is focused on the Compute elements of the OpenStack infrastructure. Future plans are to continue to enhance and extend our support for OpenStack, including focusing on networking, block storage, and the web portal.
Zenoss OpenStack ZenPack Comparison
The following table shows the key capabilities included in our existing OpenStack Cloud Monitor and new OpenStack Infrastructure ZenPacks:
[table id=1 /]
The following image shows the OpenStack component view provided by the OpenStack Infrastructure ZenPack, which displays the relationships between the various OpenStack components that make up an OpenStack device. You can see the OpenStack regions, availability zones, and other components all the way down to the hypervisors. This allows you to quickly see the relationships between all of the components for an OpenStack device.
The following image shows an example of a Service Impact model provided by the OpenStack Infrastructure ZenPack. Here you can see all of the OpenStack instances, or hypervisors, that support the tenant.
Additional Details About the New OpenStack Infrastructure ZenPack
In terms of discovery, OpenStack Nova, the compute component of OpenStack, runs a variety of different processes on the various hosts that make up your OpenStack environment. The new OpenStack Infrastructure ZenPack can show you where all the various processes such as nova-compute are running, where they are supposed to be running, and whether they are working right now.
In terms of performance monitoring, the new provider-oriented OpenStack Infrastructure ZenPack leverages and builds upon OpenStack Ceilometer, the metering and performance project within OpenStack. Zenoss is focused on leveraging the capabilities emerging within OpenStack Ceilometer for performance monitoring as much as possible, as Ceilometer is where OpenStack stores its metering and performance data. Currently, anything that is published in Ceilometer can be automatically consumed by Zenoss, so as new meters and interesting performance metrics are added into OpenStack, either as a part of an OpenStack project or by members of your team, Zenoss can consume and graph the new metrics. Zenoss also exports metering and performance data directly from Ceilometer into Zenoss in order to provide a higher level of performance than what OpenStack Ceilometer can currently provide natively.
Next Steps
Start decreasing your frustration with OpenStack monitoring today by downloading and installing the OpenStack ZenPack that best meets your needs -- either the new, provider-oriented OpenStack Infrastructure ZenPack, or our recently updated OpenStack Cloud Monitor ZenPack.
Configuring these ZenPacks is very straightforward. Simply specify the instance you want to monitor and appropriate administrative credentials for the instance. Once you supply this information, Zenoss automatically begins the discovery process and starts to monitor all of the different OpenStack components you have in your environment.
Because of the way the OpenStack Infrastructure ZenPack integrates with Ceilometer, there is also an additional step of configuring an OpenStack Infrastructure dispatcher plugin. This plugin integrates using an extension facility that already exists within Celiometer and is needed in order to address some known performance issues with Ceilometer. The dispatcher plugin ships data directly from Ceilometer into Zenoss, which means not only faster access to metrics, but also a more scalable long-term data storage solution than what Ceilometer provides natively, with no negative impact to OpenStack or Ceilometer performance.
For more information about installing and configuring the OpenStack Infrastructure ZenPack, see the detailed technical information provided on our OpenStack Infrastructure ZenPack wiki page.
Want to Know More About Zenoss?
If you are new to Zenoss, check out the following links, which provide more information about Zenoss and the unified monitoring capabilities Zenoss provides:
- Watch a 3-minute demo on Unified Monitoring with Zenoss »
See how Zenoss delivers centralized, real-time event management and complete visibility for your entire dynamic and static datacenter environments, with an extraordinarily broad list of support for operating systems, applications, hardware, environmental controls, virtualization, and cloud computing. - Read Zenoss Service Dynamics: 4 Profiles in Unified Monitoring Success »
Learn how four different companies are using Zenoss Service Dynamics today to improve their efficiency and productivity and avoid outages. - Request a free evaluation – Try Zenoss Now »
Try out all of the features of Zenoss Service Dynamics inside your own infrastructure along with full support from our world-class support team.
