Integrating with BMC Helix Intelligent Automation to send resizing recommendations for overallocated containers

On the Kubernetes cluster, create credentials for the Kubernetes cluster by using the CLI utility.
Make sure that the Kubernetes bearer token has sufficient permissions to allow resources to be resized on the Kubernetes containers.

Configuring the Kubernetes cloud connector

Configuring the Kubernetes on-premises connector

Create an automation policy with the event_type as OVERALLOCATED_CONTAINER_RECOMMENDATION. Paste the following sample JSON payload in the event.

{
"server": "",
"tags": {
"event_type": "OVERALLOCATED_CONTAINER_RECOMMENDATION",
"platform": "kubernetes"
},
"payload": {
"generated_by": {
"product": "Optimize",
"version": "21.3.00"
},
"resources": [
{
"resource_name": "",
"cluster": "",
"namespace": "",
"deployment": "",
"container_cpu_recommendation": {
"action": "reduce_cpu_allocation",
"recommended_cpu_request": "",
"recommended_cpu_limit": ""
},
"container_memory_recommendation": {
"action": "reduce_memory_allocation",
"recommended_memory_request": "",
"recommended_memory_limit": ""
}
}
]
}
}

For detailed steps, see  Creating automation policies .

BMC Helix Continuous Optimization sends an event to BMC Helix Intelligent Automation to resize the selected containers.

When you are sure about implementing the Kubernetes Overallocated Container recommendations, you can send the details to BMC Helix Intelligent Automation for automatic implementation of the recommendations. 

If the event matches a trigger in the policy, the containers are moved to BMC Helix Intelligent Automation. For details, see Recommendations page in the Kubernetes view.

• In BMC Helix Intelligent Automation, navigate to History > Policy runs page. For more information, see  Viewing automation policy runs .
• If you have access to the Kubernetes cluster, check the CPU and memory limits and requests by using either the Kubernetes Cluster UI or the kubectl command. For example:
  --kubeconfig $YAML_FILE -n <Namespace> describe deployment <Deployment> | grep -e Name -e Limits -e cpu: -e memory: -e Requests