vROps Ping Capabilities

For many years, customers were asking for up/down monitoring capabilities from VMware vROps. To a certain extent, vROps was able to provide insight into availability based on health, properties, and metrics. With 8.2, and the Ping management pack, vROps is now able to provide up/down status of targets based on ping metrics. I wrote a blog about this several months ago: https://bpatvmware.wixsite.com/bpatvmware/post/vmware-vrealize-ping-management-pack. This blog explored the Ping management pack itself: the ping adapter, the metrics it collects, dashboards, views, and alerts.

Shortly after writing that blog, I noticed a Ping Statistics section on the Summary page of a VM and thought: ah cool, they included ping statistics here for us.

After digging a little deeper though, while these are ping response metrics, they aren't being provided by the ping management pack. They are being provided by ping capabilities from the vCenter adapter itself, which was also introduced in vROps 8.2. Let's explore.

There are two different ways to monitor up/down status of objects in vROps: one via the Ping management pack and the ping adapter, the other via the vSphere management pack and the vCenter adapter. It looks like this.

The ping adapter can ping vSphere objects, non-vSphere objects like physical servers, and internet objects like URLs. The target objects are either IPs or FQDNs and are descendants of the Ping World object in vROps.

I configured a ping adapter instance to ping several network segments and a few URLs. I took all of the advanced settings defaults, but with the ping adapter we do have the ability to adjust wait times, batch size, packet size, and more. This is what it looked like.

I then cloned the Ping Overview dashboard that comes with the Ping management pack and added two heatmaps: one for IPs and one for FQDNs. These are the objects discovered by the ping adapter.

We now have a dashboard showing detailed ping metrics and up/down (green/red) objects defined by ping packet loss. If you explore these objects in more detail, you'll note that ping statistics captured by the ping adapter are available via the Metrics tab for each object.

Another way to ping objects in vROps 8.2+, is to use the ping capability from the vCenter adapter. These are the ping response metrics you see on the Summary page of the vSphere object, in this case a VM.

There are a couple different ways to enable vSphere ping, the first is via the object Summary page and the Enable Ping Monitoring option from the Ping Statistics widget.

This gives you the ability to turn on/off vSphere ping at an object level, but what if you want to turn on vSphere ping for all VMs, all ESXi Hosts, or any other vSphere object type? Do this via Administration - Inventory. Select the objects, object types, whatever objects you want to ping, then click the pencil/edit icon. Here I've selected all VMs in my lab.

Once you have the objects you want, click the pencil/edit icon and you'll be given the edit objects option, which will apply to all objects you've selected.

In this case, I've set Enable Ping monitoring to true, which enables ping monitoring against the group of selected objects (in my case all VMs) from the vCenter adapter. You can also turn the Collection Interval down to 1 minute if you wish, the default is 5.

You will now see Ping Statistics on the object summary page. Note: the vSphere ping mechanism will ping all IPs/NICs available for an object. For example, if you're pinging an ESXi Host with three IPs/NICs, it'll look like this.

When creating alerts for objects that are down, this should be considered. You can also see metrics captured by the vCenter adapter ping mechanism, they are in the Metrics - Ping Statistics folder for an object.

It's important to note objects pinged by the vCenter adapter are descendants of vSphere World, while objects pinged by the Ping adapter are descendants of Ping World.