This is where vRealize Operations Policies come into play! It gives us the ability to provide policy controls for specific service levels for various application workloads.
To start we are going to click on the Administration Link and then Policies.
When you open up the Active Policies, it shows all the policies that have associated objects assigned. Initially, there will only be the vSphere Solution's Default policy, which shows the priority rating of D. As you add additional policies, it will show the priority ranking. In the image below, we recognize the Gold SLA is priority 1, Silver SLA is priority 2, Bronze SLA is priority 3, Production is priority 4, and Development is Priority 5. If an object is a member of multiple policies, it uses these rankings to break the tie. As we will see shortly, my Gold SLA is nested in the Production policy, because of the priority ranking 1 all objects assigned to the Gold SLA policy will use this policy.
You can change the ranking of the policies simply by dragging them into the respective priority slots.
You will want to use policies with custom groups, custom groups is outside the scope of today's post, but if we look at my Environment custom groups I have organized my development, production, production bronze, production silver, and production gold virtual machines and applied these custom policies to the groups. This is the sweet spot for policies, like we discussed above we cannot treat all IT datacenter systems the same, some systems are more critical to the operation than others.
To create a new policy, we are going to click on Policy Library tab and the click the green + sign. For my example, I am going to build my policies off the vSphere Solution's Default Policy.
Let's walk through the build process for my Bronze SLA policy. First, I am going to give my new policy a name and a description, then select to Start With vSphere Solution's Default Policy.
In most circumstances, you will not make changes to the Select Base Policies, but if you wanted to modify the settings from the configuration wizard during the original installation, you can select to apply a new policy from the Overide settings from additional policies: option. In my case, since this is a bronze level SLA, I change it to VMware Test and Dev Policy (w/o Allocation).
For Analysis Settings, I am going to scroll down to the vCenter Adapter, my Workload policy has already been modified due to the override settings I specified on the base policy. I notice that my orange Workload Score Threshold is OFF, I want to change this so that I get notified at 95%.
I drag my yellow Workload Score Threshold to 90%, then double click my orange threshold and drag it to 95%.
Clicking the lock icon next to a given policy enables you to edit the value, which overrides the default setting from the base policy. You select lock to inherit settings from the parent policy.
In Time Range, I am going to unlock the policy and modify the Track Usage to only the work week by unchecking Saturday and Sunday.
You can set the workload automation options for your policy, so that vRealize Operations Manager can balance the workload in your environment. Rebalancing is an Action option available at the datacenter and cluster level to balance workloads across your hosts and clusters.
For Balance Workloads, select Aggressive balancing when you have stable populations. It minimizes contention but moves workloads more, which can cause disruption. Consider Conservative balancing when you have dynamic populations. It exposes potential contention, but moves workloads less.
For Consolidation Workloads, select more consolidation when you have virtual machines with steady demand. It puts workloads into as few hosts as possible to reduce licensing and power costs. However, this might cause less responsive capacity. Choose less consolidation when you have virtual machines with irregular demand. It uses all available hosts, which leaves more room for demand spikes. However, this increase licensing and power costs.
For Balance Workloads, select Aggressive balancing when you have stable populations. It minimizes contention but moves workloads more, which can cause disruption. Consider Conservative balancing when you have dynamic populations. It exposes potential contention, but moves workloads less.
For Consolidation Workloads, select more consolidation when you have virtual machines with steady demand. It puts workloads into as few hosts as possible to reduce licensing and power costs. However, this might cause less responsive capacity. Choose less consolidation when you have virtual machines with irregular demand. It uses all available hosts, which leaves more room for demand spikes. However, this increase licensing and power costs.
For each object type, you can override the attributes that vRealize Operations tracks. Select Enabled next to each item you want to track, Disabled for items you do not want to track, or elect Inherited to use the settings from the base policy. Like Select Base Policies, this is a setting you will rarely need to change, but it does give you the ability to become very granular when creating your policies
vRealize Operations uses alert definitions and symptoms to identify problems. Like Collect Metrics and Properties, you can override which alert definitions and symptoms will be used. Select Enabled next to each item you want to track, Disabled for items you do not want to track, or elect Inherited to use the settings from the base policy.
Profiles show you how many virtual machines will fit with the capacity remaining. By default you get four profiles, which include average, large, medium, and small. You can add new profiles on the Capacity Remaining tab under Analysis.
Like the previous settings, you can choose Enabled next to each item you want to track, Disabled for items you do not want to track, or elect Inherited to use the settings from the base policy.
To finish up my policy, I am going to select my Production Bronze Virtual Machines group and click Save.
For my gold and silver virtual machine policies, I am going to step through the same settings; however I am not going to use VMware Test and Dev Policy (w/o Allocation) in the Select Base Policies. Also, on the Analysis Settings, I am going to lower the Badge Score Thresholds for Workload and Anomaly, so that I get notified earlier in the event of a potential resource constraint.
To do this, I am going to add the vCenter Adapter – Host System and Virtual Machine, then switch on the individual analysis types to override the settings. I adjust the bars for yellow, orange, and red to be more conservative in my analysis.
After my policies are created, I can highlight it in the Active Policies window and select the Complete Settings Including Inherited tab to view all the details for the policy.
You can create new child policies underneath policies you previously produced, in the image below you will notice I created three nested policies under the Production policy, they are Bronze SLA, Gold SLA, and Silver SLA. The Silver SLA policy will inherit settings from the Production and vSphere Solution's Default Policy, and will override any settings that I modify.
