Article Written by David Marshall
For years, VMblog has covered numerous technologies coming out of the SolarWinds think tank. Perhaps the best known or most covered technology thus far (for obvious reasons) has been SolarWinds Virtualization Manager or VMAN -- the company's popular VMware vSphere and Microsoft Hyper-V monitoring and management solution.
[ Watch the SolarWinds VMAN video interview from VMworld 2016 ]
Beyond Virtualization Manager, SolarWinds has been working hard to up-level and incorporate each of the company's individual product lines into a unified view and centralized management platform by leveraging its modular, extensible and unified Orion Platform. Bringing numerous monitoring and management applications under a single umbrella was probably hard enough, but SolarWinds is about to launch something that takes the Orion Platform to a whole... other... level. And that, ladies and gentlemen, is where things get REALLY interesting. This upcoming technology is known as "PerfStack."
SolarWinds PerfStack and Orion
If you're already a SolarWinds Orion Platform owner, congrats! You're in luck. All of the SolarWinds products that currently operate within Orion will automatically be getting the upcoming PerfStack functionality face-lift. And those same individual products will each contribute data back to PerfStack -- further adding to the power of what this upcoming technology will ultimately bring to the table.
PerfStack is a new dashboard for SolarWinds Orion customers that will usher in a new way for administrators to visualize and correlate IT monitoring data. The idea here is pretty straightforward - it's a design framework that takes something extremely complicated and cumbersome, and turns it into something simple and actionable.
The new technology will combine user selected data metrics taken from across a number of SolarWinds product lines (such as Virtualization Manager, Storage Resource Manager, Network Performance Monitor, Web Performance Monitor and Server & Applications Monitor), and bring them together into one easy-to-consume data visualization, matching them up by time. Doing so provides a powerful way of improving troubleshooting techniques for countless performance issues across infrastructure and applications alike. It also allows users to quickly sift through the massive amount of data being collected by Orion, eliminating the noise and allowing the user to focus on what's truly relevant to the issue at hand.
What I initially liked about PerfStack was its highly customizable and easy-to-use interface. Data correlation and troubleshooting can be an extremely complicated undertaking, so it's important that the complexity of the process be removed as much as possible without diluting the capabilities themselves. So far, it looks like SolarWinds may have cracked that nut.
With PerfStack, IT administrators can selectively drag-and-drop historical data from any of the SolarWinds Orion platform-based products to visualize the relationship between those suspect elements. With this new visualization method, SolarWinds users can now identify, isolate and troubleshoot performance problems in ways they couldn't before. Data comparisons and correlation can be made across switches, routers, virtual machines, host servers, applications, storage LUNs, arrays, and Web transactions. Comparisons can also be made across disparate data.
For the visualization, SolarWinds relies heavily on the UI framework, specifically two recent additions: Apollo, which is an Angular JS framework, and D3, a JS graphing library. The Apollo Angular framework renders the Metric Palette, enabling succinct navigation and coupling of statistics to graphs, while D3 provides the high resolution graphs and charts.
Troubleshooting with collaboration
Another interesting (and perhaps the most powerful) element of PerfStack is the ability to collaborate with others in the organization. PerfStack can help users and teams across silos break down barriers and get on the same page so they can better work with one another while trying to troubleshoot and triage problems across functional areas. And it can help to eliminate the finger pointing battles that take place between groups with different roles and areas of responsibility -- "It's the network!" "No, it's the storage!" "It's not me, it must be your virtualization layer!".
A tool like PerfStack can provide an organization with a better holistic view of the business, but what it can't do on its own is alter corporate IT culture. While the data can now be correlated and compared across silos or across point solutions designed for a particular group, it shouldn't just get tossed in like feed to chickens. This new capability might best be introduced slowly and properly explained so that teams don't take it as an infringement against their actual or perceived controls.
Sharing made easy
Sharing analysis from PerfStack is as simple as copying the URL from a browser and pasting it elsewhere like in an email, an instant message or even a helpdesk or QA ticket. Once copied and submitted, that same URL will provide others with quick and easy access to exactly what was found during the initial troubleshooting session, eliminating questions of "how did you find this?" "what did you do?" or "what did you find?".
Even better, other users can take that same information and continue to interrogate the system by adding their own metrics without affecting the original project.
Don't worry, Orion account restrictions are still in full effect. If a user tries to open a shared or saved project and doesn't have authority or access rights to that data information or those objects, any and all metrics related to those restricted elements will be automatically hidden from the user.
Search, Drag and Drop - Let's Begin
When you first enter a new analysis project in PerfStack, you'll come to an interactive Metric Palette that has all the data points available from Orion and you'll need to start by adding at least one entity (more than likely starting off with the entity exhibiting the symptom).
You search for an entity on the left panel, browse through all of the data SolarWinds has about that entity, and then via drag-and-drop, add it onto the visualization pane on the right.
Examples always help me understand better. Let's jump in and take a look at a couple of different scenarios.
This first scenario involves an issue reported by users in the East Datacenter who cannot access the asset collect website running on a virtual server (WESTWEB01v).
The network admin is trying to determine if this is a network related issue, so they might start out by collecting the following metrics from the router (EAST-2821):
- Average Percent Memory Used / Average CPU Load
Find Related Elements / Expand Interface
Gig0/0 (TOR) [Interface]
- Percent Discard / Percent Errors
Gig0/0.100 (MPLS) [Interface]
- Percent Transmit / Percent Receive
The Network admin might then add the server hosting the Website for comparison; at which point, it is evident that this is a server issue and the sysadmin can add/remove metrics to "rule out" other issues.
The following metrics are tracked.
- Response Time
-- Average Response Time
-- Percent Loss
-- Average CPU Load (Should be spiked)
-- Average Percent Memory Used
In this next example, you'll see that the Max IOPS and Max Latency of the Virtual Datastore is being compared to the Average Latency that's reported by the VM (perfstack-vman).
And finally, we're comparing the Events, Alerts for the VM (easthyv01A), Max Latency & IOPS for the H: drive, and IO Latency Total for the Pure Storage array.
Want to get your own sneak peek? Register for this upcoming PerfStack Livecast from SolarWinds and join Head Geek Patrick Hubbard and VP Gerardo Dada on March 2, 2017 @ 11AM CST.
PerfStack is being made available to current SolarWinds customers. They can start by downloading the release candidates of SolarWinds Network Performance Monitor 12.1, SolarWinds Server & Application Monitor 6.4, SolarWinds NetFlow Traffic Analyzer 4.2.2, SolarWinds Virtualization Manager 7.1, SolarWinds Storage Resource Monitor 6.4, and SolarWinds Web Performance Monitor 2.2.1.
About the Author
David Marshall is an industry recognized virtualization and cloud computing expert, a seven time recipient of the VMware vExpert distinction, and has been heavily involved in the industry for the past 16 years. To help solve industry challenges, he co-founded and helped start several successful virtualization software companies such as ProTier, Surgient, Hyper9 and Vertiscale. He also spent a few years transforming desktop virtualization while at Virtual Bridges.
David is also a co-author of two very popular server virtualization books: "Advanced Server Virtualization: VMware and Microsoft Platforms in the Virtual Data Center" and "VMware ESX Essentials in the Virtual Data Center" and the Technical Editor on Wiley's "Virtualization for Dummies" and "VMware VI3 for Dummies" books. David also authored countless articles for a number of well known technical magazines, including: InfoWorld, Virtual-Strategy and TechTarget. In 2004, he founded the oldest independent virtualization and cloud computing news site, VMblog.com, which he still operates today.
Follow David Marshall