Scale a Running TNC-O

The following guide explains how to upgrade a running Telco Network Cloud Orchestration (TNC-O) to scale with CPU usage.

Pre-requisites:

• The Kubernetes metrics server must be running in your Kubernetes environment for scaling of TNC-O to work.

Configure TNC-O

Existing installations of TNC-O can be scaled by adding an Horizontal Pod Autoscaler via kubectl.
For each service that you want to scale do the following:

1. Ensure that the service has resources set, you can do this by running:

kubectl describe deployment <service name>

If the deployment has resources set then you will see the following:

Containers:
  <service name>:
    ...
    Limits:
      cpu:     <maximum CPU to allocate to pod>
      memory:  <maximum memory to allocate to pod>
    Requests:
      cpu:      <initial CPU to allocate to pod>
      memory:   <initial memory to allocate to pod>

2. If there are Limits and Requests for your chosen service then skip to Step 3 otherwise in order to add these to a service you can add the following to a custom values file:

<service name>:
  app:
    resources: 
      limits:
        cpu: <maximum CPU to allocate to pod>
        memory: <maximum memory to allocate to pod>
      requests:
        cpu: <initial CPU to allocate to pod>
        memory: <initial memory to allocate to pod>

For example if you wanted to scale Apollo your custom values would look as follows:

apollo:
  app:
    resources: 
      limits:
        cpu: 500m
        memory: 4Gi
      requests:
        cpu: 200m
        memory: 2Gi

And do a helm upgrade:

helm upgrade lm <helm chart> --values <custom values file>

3. Once your service has resources then you can create an autoscaler using a command like this:

kubectl autoscale deployment <TNC-O service deployment> --min=<minimum pods> --max=<maximum pods> --cpu-percent=<cpu target percentage>

For example if you want to scale Apollo between 1 and 3 pods whenever a pod exceeds 80% CPU you would use this command:

kubectl autoscale deployment apollo --min=1 --max=3 --cpu-percent=80

4. You can inspect your Horizontal Pod Autoscalers by running:

kubectl get hpa