Planning your deployment

To ensure that your IBM® Cloud App Management deployment is successful, planning is critical. A successful deployment can be completed in a few main steps. Ensure you complete all the procedures in each step. This planning deployment scenario assumes that you are an administrator working on a Linux 64-bit system.
Complete the following steps:

Step 1: Determine your hardware requirements

Do you plan to install <100 monitored resources?
Use a small demonstration (trial or proof of concept) environment. You will need the following resources: 1 VMs, 8 CPU, 32 GB memory, Disk space 100 GB.
Do you plan to install >100 monitored resources?
The required hardware resources will vary depending on metrics per minute. Determine the metrics per minute value:
  1. In the estimation spreadsheet <link>, enter an estimate of the number of monitored resources you plan to install, a metrics per minute value is returned.
  2. Based on the returned metrics per minute value, review the Planning hardware and sizing topic to determine the required number of VMs, CPU, memory and disk space.

Step 2: Determine the storage type to use

You will use the prepare-pv.sh script to create the storage class and PVs for your statefulset services. You must create the PVs for the Cassandra, Kafka, ZooKeeper, CouchDB, and Datalayer statefulset services in your deployment. During the Cloud App Management server install, when you run the prepare-pv.sh script, you must select your storage type either local or vSphere.
Local storage
Local storage uses local persistent volume storage. Local storage is recommended. Local storage is similar to hostpath, in that it creates a PersistentVolume (PV) that uses a local directory on a system. Local storage differs from hostpath because of its affinity to lock storage to the node and the local directory. This affinity prevents the statefulset pod being moved to another system and losing its storage. With hostPath, if the pod is moved, you can lose persistence. Local storage offers better performance than network-based storage like NFS or GlusterFS. With local storage, if the node or its storage gets lost, then the PV data is also lost.
For local storage, the IP is the address of the IBM Cloud Private worker node that you want to assign to a specific PV. The PV and any service, which is claiming that particular node, is permanently locked to that node. If you lose that worker node, you lose that PV and the service. If your IBM Cloud Private cluster is running on a VMware deployment with administrator access to vSphere storage, VMware drives are automatically provisioned and locally attached to the correct VM. If the VM fails, the drive is moved to a new VM with the statefulset pod. For more information about setting up vSphere storage in IBM Cloud Private, see the vSphere Cloud Provider topic.
vSphere
vSphere storage uses vSphere provisioned storage. vSphere requires existing vSphere storage class. If your IBM Cloud Private cluster is running on a VMware deployment with administrator access to vSphere storage, VMware drives are automatically provisioned and locally attached to the correct VM. If the VM fails, the drive is moved to a new VM with the statefulset pod. For more information about setting up vSphere storage in IBM Cloud Private, see the vSphere Cloud Provider topic.

Step 3: Deploy IBM Cloud Private Enterprise V3.2.1

For the preparation steps that you must complete before you install IBM Cloud Private Enterprise and the instructions to download and install IBM Cloud Private Enterprise, see the Deploying IBM Cloud Private section.

Step 4: Download and deploy the Cloud App Management server

Complete the procedures in the Installing IBM Cloud App Management on IBM Cloud Private section.

Step 5: Access the Cloud App Management console

Complete the steps in the Starting the Cloud App Management UI topic.

Step 6: Deploy the agents and data collectors

Do you have IBM Tivoli® Monitoring agents or IBM Tivoli Composite Application Manager agents (referred to as V6 agents) connecting to the Tivoli Enterprise Monitoring Server?
Configure these agents to connect to the Cloud App Management server. You can then view monitoring data on the Cloud App Management console. Complete the procedures in the Integrating with IBM Tivoli Monitoring agents section.
Do you have Cloud APM V8.1.4 agents (referred to as V8 agents) connecting to the Cloud APM server?
Configure these agents to connect to the Cloud App Management server. You can then view monitoring data on the Cloud App Management console. Complete the procedures in the Integrating with Cloud APM, Private agents section.
Do you have an environment with no previous V6 or V8 agents installed?
Complete the procedures in the Deploying ICAM Agents topics.
For configuring ICAM Data Collectors, complete the procedures in the Deploying ICAM Data Collectors topics.