Since we are in the midst of VMworld season, let’s discuss our next-generation metering solution for VMware Cloud Providers, vCloud Usage Meter 4.1.
Introduction to vCloud Usage Meter 4.1
vCloud Usage Meter 4.1 is a multi-year investment on developing a scalable, provider-grade metering appliance that meets the terms of the VMware Cloud Provider Program agreement. The main goal of Usage Meter (UM) is for Cloud Provider to VMware reporting on a monthly basis.
I’m sure many people are going to ask – what’s the difference between 3.6.1 and 4.1 of Usage Meter? Let’s start off with a diagram.
As you can see, 4.1 was built with direct vCloud Usage Insight (VUI) integration. This provides a single controller to many one-to-many UM instances for a provider.
While 3.6.1 and 4.1 will interoperate (with considerations), this new metering solution provides an easier platform to manage. This will be an ongoing evolution just like any other software solution, but will be far easier to maintain for VMware Cloud Providers (due to having a SaaS component that provides the metering logic). I’d like to highlight the following:
- Usage Meter 4.1 uses the Photon platform. This provides a secure and scalable end state. Due to the hardened nature of Photon, this will meet many of the industry standard and regulatory compliance requirements. Moreover, in-place upgrades will be built into this platform.
- Cloud-Native Reporting – 4.1 requires VUI reporting: this enables providers to complete their consumption reporting with ease and minimizing the time they spend monthly.
- The Flex VCPP Bundle is also natively reported without bundle translations.
- Future-Proof Model – in-place upgrades and dynamic updates in mind. Providers should be able to adopt new licensing, bundling, or product changes with ease. Much thought was put into how we can provide code updates without a new appliance deployment/migration and provide detailed reporting on an ongoing basis. This will help us within the program add new products in a very dynamic fashion.
As of this publishing, UM 4.1 meters the following:
- vRealize Operations
As of today, VMware vCloud Director (vCD) is not metered with UM 4.1. While this is included in the new Flex bundle, there might be scenarios where one of the Advanced SP bundles are used due to vCD usage.
Below is a table that depicts current state of UM 4.1 and the different bundling situations.
With any new release, there are always considerations.
To start, this release is not on parity or functionally equivalent to Usage Meter version 3.6.1. This will be an evolution and a phased approach like any other software solution. However, UM 4.x architecture will be able to evolve faster due to the cloud-based architecture.
- Usage Meter 4.1 can only operate with Usage Insight access. Since Usage Insight does all of the calculations for monthly metering, we must have connectivity. “Offline” reporting is a work in progress.
- If you want to utilize UM 4.1 and are using vSAN, SRM, vRA, or HDaaS today, please work with your VMware VCPP field team or aggregator on how you can manually report usage. The goal is to be on-parity with 3.6.1 in the short-term. I have no doubt with this architecture that this will be possible soon.
- NSX-V is only compatible today. NSX-T will be brought into the mix at a later time.
- VCPP Flex Bundle must be utilized on your contract. If you are on the legacy bundles, continue to use 3.6.1.
- Most important, Usage Meter is not replacing the vRealize Operations Tenant App. These solutions will continue to evolve where the Tenant App is intended for customer-facing showback/chargeback functionality.
What I’m excited about is the ability to adopt new VCPP changes on demand. Think of Usage Insight as our master controller. Each 4.1 instance can adopt new changes on demand with ease. The previous version of UM operated independently without ‘true’ knowledge of Usage Insight.
As of this writing (September 2019), if you’re using a solution over vCenter, NSX, and vROps, I’d have a hard time saying go with UM 4.1 – this adds more operational overhead for manual reporting.
Let’s walk through the deployment process.
Usage Meter 4.1 Deployment
Deployment of UM 4.1 is very straight forward and intuitive. I’m going to break this out into sub-sections.
- Deployment of OVF
- Bring up of UM 4.1 Instance
- Acceptance of Data Guidelines
- Registration to Usage Insight
- Configuration of Endpoints
Usage Meter is shipped in a single OVF just like past versions. It is a single virtual appliance. Many of our past best practices still apply such as deploy Usage Meter closest to the metered endpoints, register it for DNS, etc.
One consideration – everyone will be based off of UTC time for now.
To start off, deploy the UM OVF and select your resource path –
We can see that this is Usage Meter 4.1 (do note I am deploying a pre-release build, so descriptions/build numbers might differ).
Select your network along with select the IP Allocation. Typically, I see static IP addresses applied to UM instances.
Under Customize Template, here’s where we put the following:
- DNS Server(s)
- Domain Name of the UM instance (just the short qualified)
- DNS Search Path
- Network IP Address
- Three sets of passwords – root, umauditor, and usagemeter (this is the account you will use to log into the UI).
Input your respective sections in each of the fields.
Let’s confirm and hit the Finish button –
Once deployed, we can see the Usage Meter appliance is operational and ready for initial login.
Initial Usage Meter Configuration
Let’s log in with our “usagemeter” account and the password set inside of the OVF deployment.
First screen you’ll see is to accept the Data Usage agreement. As stated before, for UM 4.1 to operate, we must be connected to VMware and send the reporting data for aggregation.
Now, we need to test connectivity to VMware. For my lab environment, I can route externally to connect to VMware vCloud Usage Insight.
Let’s hit “Test Connectivity” –
Before we can continue, we need to register this new UM instance.
vCloud Usage Insight Registration
For successful metering, the new UM 4.1 appliance must be registered to your MyVMware account that is entitled to vCloud Usage Insight (VUI).
For my lab environment, I have a test account that I am able to register this new appliance for metering.
As we can see above, I need to register this UUID to my Usage Insight account. Let’s open up Usage Insight –
We are now going to register a new UM instance. I am selecting Trial Mode since this is a lab environment (remember, Trial mode aggregates the data, but does not submit/redirect your metered results to the Business Portal) –
We will see a new screen requesting a label, Instance ID (or UUID), Contract, and the mode of operation. Again, since I am in a lab environment, I’ll be selecting Test.
Once registered, we will see a message how we want to meter vROps or NSX. This allows us to control the logic of NSX/vROps reporting – essentially if we want to report this as standalone/isolated from a bundle.
Okay, let’s go back to my newly registered UM instance and attempt a Test Connectivity again…
Now are ready to add vCenter, NSX, and vROps.
vCenter/vSphere Metering Configuration
Under the Products subtab, we will see vCenter Server as an available option. Click the Add button to bring up the dialog –
As you can see, very similar to 3.6.1. We can put in our FQDN along with the SSO username/password while using an external PSC if required.
Once a connection is established, we must accept the certificate –
Once the certificate has been accepted, we will see the status turn to green along with an “OK” message. Awesome!
Add other vCenters by duplicating these steps.
NSX Metering Configuration
Adding NSX Manager is very similar to adding a vCenter endpoint. However, we need to tie it to the correlated vCenter. Press the Add button from the NSX sub-tab –
Accept the certificate and once it’s fully authenticated, we will see the green circle. Ready to move onto vROps!
vRealize Operations (vROps) Metering Configuration
If we have a vROps instance, this will be detected based off of the vCenter registration. In my lab, we can see my vROps instance has been detected –
As before, we need to “activate” the vROps instance, press the Edit button and let’s type in our credentials –
We are now prompted for the certificate acceptance. Once completed, we should see vROps as healthy.
All notifications are done via the Notifications tab at the top. From here, one can view any alerts.
We can also change the logging level –
Usage Insight Reporting
Once the new UM is sending data to Usage Insight, one can log in and look at the current state of the Monthly Usage Report along with downloading a copy of the VM History Report –
If we take a look at the VM History Report, it looks very similar to what we’ve seen with 3.6.1, with a few changes:
- The Name, Guest Hostname, VHostName, and ESXi Host Name are all hashed. This is to strengthen our Personally Identifiable Information (PII) policy and to show that we do not collect this type of data.
- From what was explained to me, this happens by the UM instance – it is the only instance that encrypts it before transmission.
- We can see some of the columns have moved around – vSphere License has been moved to the left while we see vROps and NSX (along with NSXFInt) moved to the right. Same feature detection available.
We can see in my environment that I have some workloads utilizing vROps Enterprise along with NSX Advanced –
Here’s a few things I’ve learned so far working with the new Usage Meter 4.1 appliance. I will update this as I find new items.
Verifying Connectivity to Usage Insight
If we aren’t seeing reporting to Usage Insight, we can verify to see if the UM instance can access the Usage Insight API.
From the terminal, type the following:
curl -kv https://cloud.vmware.com/um/api/v1/ping
We can see if the UM instance can create a curl connection and read some of the details of it.
Checking UM Service Status
The main UM directory resides under /opt/vmware/cloudusagemetering – there are several subfolders here.
To check the status of the UM service(s), from this directory, type the following:
As we can see, all services are running, but I have an error under my NSX-V collection:
There is also start.sh, stop.sh, and status.sh available.
By default, SSH is not enabled for remote connections.
Open up the console, start SSHD using the command:
service sshd start
From there, we also want to permanently enable it for anytime the appliance reboots.
systemctl enable sshd
Note you will need to login using the “usagemeter” account to gain remote access. I do not recommend enabling root login for remote access.
Installing IPUtils / Ping
For some reason, this GA release did not have the IPutils package built in to the base Photon image.
You can install “iputils” by typing the following – assuming you have routed access to the public repo –
tndf install -y iputils
From there, we now have ping (and many other utilities) available.
Changing Network Settings Post-Deployment
The typical VAMI script works for post-deployment reconfiguration.
This is located under:
The script you will run is vami_config_net. From there, we can modify our gateway, hostname, DNS, Proxy Server, and IP Address.
Usage Meter 4.1 is intuitive and easy to deploy. While it does not cover all metered products yet, the architecture is established to add these items rapidly. I look forward to seeing how Usage Meter 4.1 progresses and the value it can deliver to our global VMware Cloud Providers.