Chapter 16. How does subscription watch show my subscription data?
Subscription watch shows subscription data for Red Hat offerings such as software products or product sets, organized by the Red Hat software portfolio options in the navigation menu. Each portfolio displays on its own subscription watch page, and the submenus filter the data on the page according to available product architectures, products, or product sets within the selected portfolio.
Currently, subscription watch shows subscription data for Red Hat Enterprise Linux and Red Hat OpenShift.
Each product page offers multiple views. These views enable you to explore different aspects about your subscriptions. When combined, the data from these views can help you recognize and mitigate subscription noncompliance problems or trends, organize subscription allocation across all of your resources, and improve decision-making for future purchasing and renewals.
16.1. How to use the subscription data in the views
The graph view, with its details about your subscription usage and capacity across the organization, helps you track usage trends and determine utilization, which is the percentage of deployed software when measured against your total subscriptions. The table view, with details about the usage on the individual systems within your inventory, helps you resolve questions you might have about usage and refine plans for future deployments on individual systems.
The table view data is derived from data in the Cloud Services platform inventory application. Access to subscription watch, inventory, and other applications is controlled independently by a role-based access control (RBAC) system for the Cloud Services platform tools. If you do not have access to the inventory application, you cannot view the current system inventory data in the table view. For more information, contact the organization administrator for your account.
The usage and utilization graph view
This view shows you your total subscription usage and capacity over time in a graph form. It provides perspective on your account’s subscription threshold, current subscription utilization, and remaining subscription capacity, along with the historical trend of your software usage. The usage and capacity calculations that appear in the graph are based on a data snapshot that is provided once every 24 hours by the Cloud Services platform processing tools.
Usage is the actual or equivalent consumption of physical hardware. The terms of subscriptions determine the physical hardware that is consumed, such as sockets or cores. Usage is represented by an area graph, with different types of usage, such as physical, virtual, and public cloud usage, represented by different colors.
Capacity is the sum of similar subscriptions across all of your contracts. Based on this sum of your subscriptions, the maximum capacity, also known as the subscription threshold, is represented by a dashed line.
Utilization is the percentage of capacity, as indicated by the subscription threshold, that is exhausted through the deployment and usage of Red Hat software in this account.
Although the graph shows trends over a selected time interval, you can also view more specific information for the graph. For example, if the selected time interval is Weekly, you can hover over the graph near a date to see more specific data for a particular week.
You can also use the available filters, which can vary by product or product set, to change the usage data that displays in the graph. For example, you can filter by the time interval or the unit of measurement, or by the subscription attribute filters such as service level agreement (SLA), as applicable.
The current systems table view
This view shows you an inventory of the systems that compose the most recent daily snapshot of data in the graph view. It provides information that can help you correlate the usage totals in the graph with the current software deployments on individual systems across your organization.
The data in the current systems table view contains aspects of the data that is available from the Cloud Services platform inventory application, with the following differences:
- The inventory application shows significantly more system data. The current systems table view is a small subset of this data.
- Data in the inventory application can be more current because of the methods that are used to update the data. The current systems table view in subscription watch is based on a daily snapshot, so that data could be up to 24 hours old.
- Consumption of sockets or cores in the inventory application is represented as actual consumption. Usage in subscription watch is represented as normalized consumption, bound by the terms of subscription. For example, usage of a physical RHEL subscription is measured by socket pair, so a socket count for that type of system is always rounded to the next higher even number.
The information in the current systems table shows the name of the system, the type of the system, the usage total for that system according to the unit of measurement, and the date that the system was last seen.
The system is the machine, either physical or virtualized, on which the product or product set is deployed. The system is usually represented by either its display name or its universally unique ID (UUID). For multi-guest systems such as hypervisors, you can expand the system to see more information about individual guests. You can also click a system name to open the full system record in the inventory application.
The infrastructure type is the type of system on which the product or product set is deployed. A system can be a physical host, hypervisor, individual virtual machine, or other form of virtual deployment such as a public cloud instance.
The usage total is the actual or equivalent amount of physical hardware that the product or product set is consuming on that system. The usage is counted according to the unit of measurement, which in turn is determined by the terms under which a subscription to a product or product set is sold. For example, a subscription might be sold by sockets or cores.
The last seen date is the date that the system was last seen by the Cloud Services platform inventory application. As part of the underlying tasks that subscription watch and other Cloud Services platform tools perform to calculate usage, the inventory application helps to identify and deduplicate system data that is gathered by the various data collection tools.
As with the usage and utilization graph, you can use the filters to change the data that displays in the current systems table. However, a change to the time interval, such as changing from days to weeks, has no effect on the current systems table. The data displayed is from the most recent daily snapshot, so it is usually no more than 24 hours old.
16.2. Measurement of usage and capacity
Currently, subscription watch tracks certain types of Red Hat Enterprise Linux and Red Hat OpenShift. The data that is displayed for usage and capacity can vary by product.
Overall usage and capacity trends display on the usage and utilization graph. The information in the current systems table provides additional detail about the most recent day of data from the graph.
Red Hat Enterprise Linux
- RHEL usage
Usage is measured in CPU sockets. Data is aggregated for all supported architectures and is divided by architecture, including the RHEL variants for x86. You can view aggregated or specific architecture data by selecting from the Red Hat Enterprise Linux options in the navigation menu.
The usage data in the graph is divided into three sections, based on RHEL on physical systems, virtualized systems, or public cloud.
- RHEL capacity
- To measure capacity, the socket contribution of each RHEL subscription is added to a total that encompasses the inventory’s CPU architecture, including the RHEL variants for x86.
For some Red Hat products, RHEL is included with and is installed to support that product. For example, RHEL is included with Red Hat Satellite. That type of RHEL is not tracked or counted against total usage or capacity.
Red Hat OpenShift
- Red Hat OpenShift usage
- Usage of Red Hat OpenShift Container Platform is measured in CPU cores or sockets. Data displays as an account-level view that is a sum of usage across active clusters.
- Red Hat OpenShift capacity
- To measure capacity, subscription watch accesses the Red Hat internal subscription services to display the current and recent historical trend of Red Hat OpenShift Container Platform subscription capacity, as measured in cores or sockets, for comparison against the usage data.
For Red Hat OpenShift Container Platform version 4.1 and later, subscription watch is able to distinguish between overhead and compute nodes, also commonly referred to as infrastructure and worker nodes. In the aggregation of usage data for these versions of Red Hat OpenShift Container Platform, infrastructure nodes are ignored.
Subscription watch is not able to make this same distinction for earlier versions of Red Hat OpenShift Container Platform, so data for infrastructure nodes is displayed and counted along with the worker node usage. Analysis of cluster data indicates that approximately 15% of data displayed for earlier versions of Red Hat OpenShift Container Platform is infrastructure node overhead. Therefore, if your subscription profile includes Red Hat OpenShift Container Platform version 3, it is possible that you can exceed your Red Hat OpenShift subscription threshold by up to 15% but still be in compliance with your subscriptions.
16.3. Units of measurement
Red Hat Enterprise Linux
Because of the inherent differences between physical, virtual, and public cloud offerings and their relation to hardware, subscription watch tracking uses different units of measurement, as follows:
- Physical usage
Subscription watch measures your physical RHEL installations by CPU socket pairs. Each system contributes its installed socket count, rounded upwards to the next even number. The value that displays is the total socket count, including all of the system-level pair rounding.
In the current systems table, on-premise physical hardware and other structures such as a RHEL based hypervisor can display as physical machines.
- Virtualized usage
Subscription watch measures your virtualized RHEL installations in two ways. Where host-guest mappings are not used, such as with standard guest subscriptions, each system contributes a single installed socket. Where host-guest mappings are required, such as with virtual data center (VDC) subscriptions or similar virtualized environments, the socket count of the hypervisor host node is counted, by using the same socket pair method that is used with physical RHEL installations.
Virtualized usage for hypervisors and virtual machines is grouped together in the usage and utilization graph, but hypervisor usage is displayed separately from virtual usage in the current systems table. This separation can help you troubleshoot questions about the collection of usage data for virtualized environments. In particular, it can help you determine whether host-guest mapping data is being correctly provided to subscription watch through the configuration of virt-who and the Satellite inventory upload plugin. For example, when these tools are correctly configured, virtualized usage is counted as follows:
- For a RHEL based hypervisor with RHEL guests, the socket count of the hypervisor is counted twice, with the socket pair method applied. One count as physical represents the node’s own copy of RHEL, and one count as virtualized represents the usage of guest systems.
- For a non RHEL based hypervisor with RHEL guests, the socket count of the hypervisor is counted once, as virtualized, with the socket pair method.
- For standalone virtual machines, or for virtual machines with no detectable hypervisor management, each virtual machine is counted as a single socket.
- Public cloud usage
Subscription watch measures public cloud RHEL installations by socket. The instances launched from public cloud RHEL images are recognized through Desktop Management Interfaces (DMI) fact-value pairs that are present in the image and instance metadata. The values of the DMI facts identify an instance as running in the cloud infrastructure provided by Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform (GCP), and Alibaba Cloud. Each instance contributes a single socket to the socket count.
Public cloud usage displays separately in the usage and utilization graph, but displays as virtual in the current systems table.
Red Hat OpenShift
Subscription watch measures your Red Hat OpenShift usage in units of CPU cores or CPU sockets. For Red Hat OpenShift 4, the counting is aggregated at the cluster level, and for Red Hat OpenShift 3, the counting is aggregated at the node level. Currently, subscription watch cannot display a single, mixed-unit view of Red Hat OpenShift usage in environments that include core-based and socket-based clusters within the same account. You must use filtering to view that data in separate views.
You can use a filter to toggle the usage and capacity data between the two units of measurement. If subscription attributes are set on the cluster (through Red Hat OpenShift Cluster Manager for Red Hat OpenShift 4) or on the node (through the command to set the
ocm.units value for Red Hat OpenShift 3), then that data can be reported by cores or sockets. If subscription attributes are not set or cannot be set, then the data is included in reports for both core- and socket-based usage.
- Physical usage
Subscription watch measures your core-based physical Red Hat OpenShift installations by actual core count. Socket-based physical installations are measured by socket pairs, so the count is rounded upwards to the next even number.
In the current systems table, an example of a physical system for Red Hat OpenShift is a Red Hat OpenShift cluster running on bare metal.
- Virtual usage
Subscription watch measures your core-based and socket-based installations by actual core and actual socket count.
In the current systems table, an example of a virtual system for Red Hat OpenShift is a cluster installed in environments such as Red Hat OpenStack Platform, Red Hat Virtualization, VMware vSphere, or on public cloud.
You can further refine the subscription watch data by selecting values from the available filters in the interface. When you select a filter option, the graph view (and in some cases, the table view) refreshes to show data that relates to that option. In other words, a filter is inclusive, not exclusive, for the selected option.
Filtering by time
You can filter data by several different time intervals, including daily (the default) and quarterly. Filtering by time affects only the usage and utilization graph view. The current systems table view is always data from the most recent subscription watch daily snapshot, and is not affected by the time filter.
During the rapid development of subscription watch, the addition of new features is improving the scope and accuracy of this tool. Subscription watch does not provide in-application capability to recalculate older usage and capacity data as these new features are being added. Therefore, the selection of a longer time interval could display results that contain inconsistencies.
Filtering by subscription attributes
You can filter by subscription attributes, which is data that describes the characteristics and intended usage of subscription. The accuracy of those filters is dependent upon how accurately the subscription attribute data is set.
Subscription attributes might be configured from the operating system or its management tools, or from settings within the product itself. In these various tools, subscription attribute data is also known as system purpose, subscription settings, or similar names. In some cases, subscription attribute values might be derived from the subscription, such as when a subscription is sold either by sockets or cores.
You can use the subscription watch filters to get a more focused view on usage that meets certain use cases within your subscription profile. For example, filtering your RHEL subscriptions by service level agreement (SLA) to show only those with an SLA of Premium could help you determine the current usage of premium subscriptions compared to your overall capacity for those premium subscriptions. In turn, this knowledge can inform decisions such as additional deployments, actions to mitigate subscription compliance issues, or future purchasing and renewals.
As another example, selecting a nonspecific value for a filter, such as the No SLA or Unspecified options, can help show subscriptions that have subscription attribute values that might be missing or that might be less common and not specifically filterable by subscription watch. For those subscriptions with missing subscription attributes, adding that data can improve the accuracy and usefulness of subscription watch reporting.
Subscription watch provides the following filters and filter options for RHEL:
- SLA (service level agreement): Premium, Standard, Self-Support, No SLA
- Usage: Development/Test, Disaster Recovery, Production, Unspecified
Subscription watch provides the following filter and filter options for Red Hat OpenShift:
- SLA (service level agreement): Premium, Standard, Self-Support, No SLA
- Cores: Cores (default), Sockets