If you have been keeping up with the Thwack product blog lately you know that Drag & Drop Answers to Your Toughest IT Questions revealed PerfStack, the new way to view performance analysis by visualizing and correlating data within the Orion web interface and Get Your Head Out of the Clouds, and Your Data Too identified that you can collect configuration and metric data directly from Amazon Web Services® (AWS), and have that data visualized along with the rest of the environment you are already monitoring in SAM 6.4 or VMAN 7.1. This is great news for any administrator that needs to troubleshoot a Hybrid IT application with on-premises VMs and AWS cloud instances.
The good news is that Virtualization Manager (VMAN) 7.1 allows you to leverage the "new hotness" found in PerfStack and Cloud Infrastructure monitoring to correlate and pinpoint where the application performance issue is in your hybrid infrastructure. In the following example, we have a hybrid cloud application that has resources in the cloud (SQL AWS) as well as an on-premises virtual server (Analytics VM) both of which are monitored with VMAN. As with most IT incidents, the administrator is left trying to figure out what exactly is causing the performance degradation in their application with little to go on other than the "application is slow". Using PerfStack you can quickly dive down into each KPI and drag and drop the metrics you want to compare until the troubleshooting discovery surfaces what the issue is or isn't. The fact that VMAN contains cloud infrastructure monitoring means that you can add AWS counters from your cloud instances into PerfStack and correlate those with other cloud instances or with your on-premises VMs to troubleshoot your hybrid infrastructure with VMAN.
In the example above, the cloud instance, SQL AWS is experiencing some spikes in CPU load but it is well within the normal operating parameters while the on-premises VM, Analytics VM is experiencing very little CPU load. With PerfStack my attention is easily drawn to the memory utilization being high on both servers that participate in the application's performance and the fact that my on-premises VM has an active alert gives tells me I need to dig into that VM further.
By adding Virtualization Manager counters that indicators of VM performance (CPU Ready, ballooning, Swapping) I see that there are no hidden performance issues from the hypervisor configuration (figure below).
From the Virtualization Manager VM details page for the Analytics VM, I see that the active alert is related to large snapshots on the VM which can be a cause of performance degradation. In addition, the VM resides on a host with an active alert for Host CPU utilization which may be a growing concern for me in the near future. To monitor hybrid cloud infrastructure in VMAN allows the administrator the ability to create a highly customized view for discovery and troubleshooting with the contextual data necessary minus the alert noise that can regularly occur.
One of the added benefits of monitoring cloud instances with VMAN is that you can now build on the single pane of glass view that will be the monitoring authority for both your on-premises and cloud environment. Not only is it essential to have one place to monitor and correlate your data when applications span across hybrid environments but having visibility into both environments from Solarwinds allows you to determine what requirements will be necessary when moving workloads into your AWS cloud or off your AWS environment.
For more information on PerfStack and Hybrid end-to-end monitoring, check out the following link.
Don't forget to check these blog posts for deeper dives into PerfStack & Cloud Monitoring respectively.