VMworld 2018 presentations

As requested, you can find the deck here. They are in PowerPoint, not PDF.

First session was Operationalize Your World, where you learn about transforming from reactive, complaint-based operations to proactive, insight-based operations. It’s using Performance SLA and KPI. You get KPI defined for VM, Cluster and Multi-Tier Applications. Below is an example of how a VM KPI is calculated.

The second session was vSphere counters deep dive. It does not duplicate information generally covered in vSphere documentations or whitepaper. For example, I explained that measuring memory is hard. MS Windows actually include non active pages in its In Use counter. This explains why vCenter VM Active RAM counter is lower.

Hope you found the 2 sessions useful!

Your vR Ops. A bespoke operations tool

As human, we love tinkering. We customize our cars, our phones, etc. That’s why toys like Lego is hugely popular. In vR Ops, we recognise that each operations is unique. 100 companies may use the same identical architecture (e.g. they all use VMware on Amazon), but the way they run operations will be unique. While they are similar, at the ground level you will have 100 different operations.

One key feature of vR Ops is customization. One cloud provider customized by building their own UI. Another example of advance customisation is the vR Ops vCD Tenant App. You get a completely different app!

Now, you don’t have to spend that much engineering effort. Within a few minutes, you can customize vR Ops. I’ll show you one example, as this is a popular one among my customers.

In the above, my customers want to create a troubleshooting flow. It enables you to drill down by simply clicking on the object name. We leverage vR Ops ability to customize

  • the Summary Page.
  • the out of the box Dashboard.

So the implementation looks like this. We customize the Troubleshoot a Cluster Dashboard, and 3 Summary Pages.

Here is what the customized Troubleshoot a Cluster dashboard. You can access it via the Getting Started, because you’re using the same dashboard ID. Cool!

You can sort by any column. You can also change the time period (yes, no need to have Edit access anymore!). If you have 100s clusters, you can also filter to specific vCenter. Yup, these are new features in vR Ops 7.0!

Once you find the cluster you need, simply click on it. It takes you the Cluster Summary Page. I explained it here.

You can see whether the performance (read: high contention) is caused by high utilization or not. You can also see if the problem is spread across multiple hosts or not. From here, you can drill down into the Host. I explain the dashboard below here.

Finally, from the host, you can drill down into a VM.

The implementation does not use Group nor Policy. It’s certainly heavy on super metric.

You can download all the above dashboards from Sample Exchange. Hope that gives you an idea to customize your vR Ops to meet your unique operations. And have fun tinkering!

VM Performance dashboard

Continuing from the vSphere Cluster Performance dashboard and ESXi Performance dashboard, here is the VMware VM Performance dashboard:

The dashboard follows the same layout with cluster and ESXi dashboards. If it’s not clear, let me know.

The dashboard implements the concept I shared here. Do read it, as most of my customers are surprised with the choice of counters. I’m not saying your counters are wrong. I’m saying you can be more accurate in certain cases.

As VM is the smallest object, the dashboard does not show any child objects. Rather, I added a Trend line chart. This helps you check if the spike is something expected.

Lastly, the dashboard is complemented with a List page. How do you know which VMs to look at first? Well, the list below should help you:

As usual, get the dashboards from VMware Sample Exchange. For instruction on how to install, read the ESXi Performance dashboard.