• Brock Peterson

The Evolution of a vROps Dashboard

Updated: Mar 8

I first heard about VMware {code} a couple years ago, since then it has become my favorite VMware-related site. It has all sorts of content, from things like Dev Centers and SDKs, to a Sample Exchange and an API Explorer. I spend most of my time in the Sample Exchange, which people use to share things like vROps dashboards and super metrics, vRA Blueprints, vRO Packages and Workflows, vRLI Content Packs, and more.

I've contributed several dashboards to VMware {code}, one of my first focused on troubleshooting VMs:

It's a simple dashboard, using only three widgets. Top left is an object list widget configured to show VMs. Top right is an advanced object list widget showing parent and child objects for the selected VM. Below them are 14 metric chart widgets designed to show CPU, memory, disk, and network metrics for the selected VM.

The user can quickly search for a VM via the filter in the object list widget and see the VMs health, related object health, and 14 key performance indicators (KPIs). While not terribly complicated, it's a useful dashboard.

A year later I wanted more information about the VM itself in the object list widget. I also wanted to know if the VM (or its parent/child objects) had any alerts, so I built Troubleshooting VMs v2:

This dashboard uses five widgets, the original object list which feeds all the others, the advanced object relationship, and the metric charts. I added the alert list widget to show alerts and three instances of the scoreboard widget to eliminate some of the metric charts. The scoreboard widget is one of the most powerful ways to publish data in vROps, giving you the ability to show multiple metrics, history, symptom based/threshold based colors, and much more.

This morning, I released the third version of this dashboard, Troubleshooting VMs v3:

It uses four widgets: view, alert list, advanced object relationship, and health chart. Similar to scoreboards, the view widget is one of the most powerful and most often used widgets in vROps. The view widget itself is just a pointer to the underlying view, another vrops construct. Views can be lists, summary lists, trend lines, distribution charts (pie, bar, or donut), text, or images. Views are often times used instead of object lists because they are much more powerful. Among other things, you can adjust units and provide color based thresholds for metrics.

The design is to show VM KPIs at the top, with color based thresholds, sortable columns, and the ability to search for a VM. The view widget allows you to do that, everything below it is driven based on the selected VM.

The alerts and relationships are similar to what we've already seen, but I've introduced more detail around disk and network KPIs. Often times, vROps will alert on disk and network usage, but you don't have much detail beyond that. The views here are designed to show exactly which filesystems/drives are filling up and which vNICs are using the most network bandwidth.

Additionally, I replaced the metric charts with health charts (pointed at KPIs) to leverage some of the newer features in the health chart widget. They provide some coloring around alert/symptom states, minimum/maximum/current values, and are generally just more robust than the metric chart widget.

I'm guessing someday there will be a fourth version of this dashboard, but until then, happy dashboarding!

302 views0 comments

Recent Posts

See All