Workload placement can be a complex problem. vRealize Automation enforces business and IT policies relating to placement. vRealize Operations can make placement recommendations based on health and capacity metrics. Beginning with vRealize Automation 7.3 and vRealize Operations 6.6, you can leverage the capabilities of both products to optimize your workload placement.
The high-level flow across the two products (and vSphere) is as follows:
vRealize Automation is responsible for Governance, or determining the eligible placement targets for a provisioning request. In this case, we are showing a request for a blueprint in a particular business group. This business group has access to a handful of reservations. Each reservation includes a vCenter cluster and one or more datastores or datastore clusters. In our example, the blueprint also has a reservation policy (represented by the tag). The same reservation policy only appears on two of the reservations available to the business group, so only those reservations are eligible for this request.
vRealize Automation passes the eligible reservations to vRealize Operations for Optimization. Based on operational metrics, vRealize Operations returns a recommendation from the options presented by vRealize Automation. You can configure a placement policy to control whether to consolidate workloads into the fewest possible clusters or balance workloads across the available clusters. You can also set the amount of headroom to leave space for organic growth or accommodate spikes in usage. Read more about workload placement policies in vRealize Operations in this post on Automated Workload Balance.
The workload placement recommendation from vRealize Operations identifies the best cluster and datastore (or datastore cluster) to provision the machine. In the final step, vSphere DRS or Storage DRS determines the specific host or datastore within the recommended cluster.
You can view the details of the recommendation from the Execution Information pane of the Request Details page:
This list shows the eligible reservations for this request according to vRealize Automation policies. We can see that vRealize Operations has selected the Finance-Gold reservation, along with one of the datastore clusters associated with that reservation, as the placement recommendation. The icon next to the Finance and Copy of Finance reservations indicates that these clusters are not managed by vRealize Operations. Since these unmanaged clusters will never be recommended, you should consider bringing them under management by vRealize Operations.
Get Started with Workload Placement
- Install or upgrade to vRealize Automation 7.3 and vRealize Operations 6.6.
- Install the Management Pack for vRealize Automation.
- Review the detailed documentation about configuring workload placement across vRealize Automation and vRealize Operations.
This initial release of workload placement is targeted for optimizing the initial placement decision made at deployment time. In future releases, we plan to expand on this functionality to additional use cases including ongoing rebalancing.