By: Mike Langdon
If you’re running Microsoft SQL Server on VMware, hosted on Cisco servers and connected through Cisco switches to NetApp Storage, pinpointing the root-cause of a problem can be a real challenge. After all, how do you know if poor performance is caused by I/O bottlenecks, switch latency or capacity issues?
In conversations with customers, I find that this is often the biggest struggle that they face — it’s nearly impossible for them to determine where issues originate, particularly as the IT stack grows in complexity. In an enterprise setting, for instance, organizations typically have hundreds of servers, switches and storage appliances, and thousands of virtual machines.
As a database administrator, you most likely know what virtual machine hosts your database — the hostname matches, but you probably don’t know what the virtual host and datastore that virtual machine lives on. And if you don’t have insight into the virtual layer, forget about knowing anything about the hardware. You’re troubleshooting through a keyhole.
vRealize Operations serves as a central console for all of your data — from hardware to software. Within vRealize Operations, you have the ability to monitor all elements of your IT stack — whether it includes SQL Server, Cisco Networking or Cisco UCS. When you pair it with management packs from Blue Medora, you can extend your monitoring to the next level — enabling access to expansive metrics and dashboards, but also giving you insight into the relationships between the pieces that make up your applications.
The management packs connect everything to the virtual layer within vRealize Operations so you can monitor your hardware, virtual machines and databases all in one tool. However, connecting them together to map key relationships from each layer can transform your troubleshooting.
Apache Hadoop is the 500 pound gorilla in distributed big data, supporting an ecosystem for processing the large datasets that organizations rely on in today’s business environment.
Having insight into key metrics like system CPU, disk capacity, containers and applications establishes a foundation for your monitoring — but as I highlighted earlier, to truly understand where your performance breakdowns occur, that needs to extend beyond just your database metrics.
The Apache Hadoop Management Pack for vRealize Operations from Blue Medora includes relationships to other key elements of your IT stack, including VMware virtual machines, and other virtual platforms like KVM.
Figure 1: The VM to Apache Hadoop Node Dashboard from the Blue Medora Management Pack
In turn, when an issue does arise, you can easily pinpoint whether the issue comes from an issue processing a data set within YARN, or if it stems from a resource contention between nodes on a VMware host. As a result, your troubleshooting times decrease significantly — creating a better experience for your end users and optimizing performance across the IT stack.
Interested in learning more? Check out our free on-demand webinar, Connect the Dots from Hardware to Software with Relationship Mapping.
This blog post first appeared on the VMware Cloud Management blog. Read the full post here.