Requirements
To activate the autoscaler within the Kubernetes cluster, make sure you have:
- access to the cluster via kubectl for verification or debugging operations (optional but recommended);
- Helm installed locally (installation guide);
- a Node Pool with autoscaling enabled;
- an active network connection to the cluster (VPN or private access if the cluster is not public);
- a valid Client Credential, consisting of a Client ID and Client Secret.
These requirements ensure that the autoscaler can be installed and communicate correctly with the infrastructure.
Where to find the required parameters
To install the autoscaler correctly:
- access the cluster details page;
- locate the box dedicated to activating the autoscaler;
- select View precompiled chart;
- the Helm chart window displays all the required information:
- specifically:
- ClientID is the Client ID of the Client Credential that you must enter in place of the
ClientID value;
- ClientSecret is the Client Secret of the Client Credential that you must enter in place of the
ClientSecret value;
- Project ID is the ID assigned to the project (the value is already filled in);
- Cluster ID is the ID assigned to the cluster (the value is already filled in);
- Region is the region where the cluster is located (the value is already filled in);
- image.tag is the Kubernetes version (the value is already filled in).
To protect the sensitive data of the Client Credential (Client ID and Client Secret), we recommend storing them as Kubernetes
Secrets in the
kube-system namespace. This allows the autoscaler to use them without entering them directly in commands or configuration files.
Installing the autoscaler with Helm chart
Checking the deployment status
After installation, you can check that the autoscaler is active and running correctly in the cluster.
Open the terminal and run:
k get po -A
In the
kube-system namespace, locate the pod
<pod-name> (with a name similar to
aca-arubacluster-autoscaler-...).
It must have the
Running status, for example:
...
kube-system <pod-name> 1/1 Running
...
To monitor its activity in real time, you can view the logs with:
k logs -f -n kube-system <pod-name>
What to do if it does not work
If, after installation, the autoscaler does not start or does not scale as expected, check the following: