Node pools and scaling in Azure - Step-by-Step Execution

Process Flow - Node pools and scaling

Create AKS Cluster

↓

Add Node Pool(s)

↓

Set Node Pool Size

↓

Monitor Load

↓

Scale Node Pool Up or Down

↓

Update Cluster State

This flow shows creating a cluster, adding node pools, setting their size, monitoring load, and scaling node pools accordingly.

Execution Sample

Azure

az aks create --resource-group myRG --name myAKS --node-count 2
az aks nodepool add --resource-group myRG --cluster-name myAKS --name np1 --node-count 3
az aks nodepool scale --resource-group myRG --cluster-name myAKS --name np1 --node-count 5

This code creates an AKS cluster with 2 nodes, adds a node pool with 3 nodes, then scales that node pool to 5 nodes.

Process Table

Step	Action	Node Pool	Node Count Before	Node Count After	Cluster State
1	Create AKS cluster	default	0	2	Cluster with default node pool of 2 nodes
2	Add node pool np1	np1	0	3	Cluster has default (2 nodes) + np1 (3 nodes)
3	Scale node pool np1	np1	3	5	Cluster has default (2 nodes) + np1 (5 nodes)
4	Monitor load	-	-	-	Load monitored, triggers scaling if needed
5	Scale node pool np1 down	np1	5	4	Cluster has default (2 nodes) + np1 (4 nodes)
6	Scale node pool np1 up	np1	4	6	Cluster has default (2 nodes) + np1 (6 nodes)
7	Exit	-	-	-	Scaling complete, cluster stable

💡 Scaling operations complete, cluster node pools adjusted as requested

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 5	After Step 6	Final
default_node_pool_count	0	2	2	2	2	2	2
np1_node_pool_count	0	0	3	5	4	6	6
cluster_total_nodes	0	2	5	7	6	8	8

Key Moments - 3 Insights

Why does the cluster have nodes before adding a new node pool?

Because the default node pool is created with the cluster initially (see Step 1 in execution_table), so nodes exist before adding extra node pools.

What happens if you scale a node pool to a smaller number than before?

The node count decreases accordingly (see Step 5), reducing cluster capacity but keeping the cluster stable.

Does scaling one node pool affect the other node pools?

No, scaling affects only the specified node pool (see Steps 3, 5, 6), other node pools keep their node counts.

Visual Quiz - 3 Questions

Look at the execution_table, what is the node count of node pool 'np1' after Step 3?

A 3

B 2

C 5

D 0

💡 Hint

Check the 'Node Count After' column for Step 3 in execution_table.

At which step does the cluster have a total of 8 nodes?

A Step 5

B Step 6

C Step 3

D Step 2

💡 Hint

Look at 'cluster_total_nodes' in variable_tracker after each step.

If you remove the default node pool, how would the cluster total nodes change after Step 1?

A It would be 0

B It would be 2

C It would be 3

D It would be 5

💡 Hint

Default node pool count is 2 after Step 1; removing it means zero nodes initially.

Concept Snapshot

Node pools are groups of nodes in an AKS cluster.
You can add multiple node pools with different sizes.
Scaling changes the number of nodes in a node pool.
Scaling one node pool does not affect others.
Monitor load to decide when to scale up or down.

Full Transcript

This visual execution shows how an Azure Kubernetes Service (AKS) cluster is created with a default node pool. Then, a new node pool named 'np1' is added with 3 nodes. The node pool 'np1' is scaled up to 5 nodes, then later scaled down and up again. The cluster state updates after each scaling operation. Variables track the number of nodes in each pool and total nodes in the cluster. Key moments clarify that the default node pool exists from the start, scaling affects only the targeted node pool, and scaling down reduces nodes safely. The quiz tests understanding of node counts after steps and effects of removing node pools. The snapshot summarizes node pools and scaling basics in AKS.

Practice

1. What is the main purpose of a node pool in an Azure Kubernetes Service (AKS) cluster?

easy

A. To manage user access and permissions

B. To group nodes with similar configurations for easier management

C. To store application data persistently

D. To monitor network traffic between pods

Solution

1. Step 1: Understand node pool concept

   A node pool is a set of machines (nodes) in a cluster that share the same configuration, such as VM size and OS.

2. Step 2: Identify the main purpose

   Grouping nodes with similar settings helps manage and scale them easily within the cluster.

3. Final Answer:

   To group nodes with similar configurations for easier management -> Option B

4. Quick Check:

   Node pool = group of similar nodes [OK]

Hint: Node pools group similar machines in a cluster [OK]

Common Mistakes:

• Confusing node pools with storage or permissions
• Thinking node pools monitor network traffic
• Assuming node pools store data

2. Which of the following is the correct Azure CLI command to scale a node pool named pool1 in an AKS cluster called myCluster to 5 nodes?

easy

A. az aks nodepool scale --resource-group myResourceGroup --cluster-name myCluster --name pool1 --node-count 5

B. az aks scale nodepool --cluster myCluster --name pool1 --count 5

C. az nodepool scale --cluster myCluster --name pool1 --nodes 5

D. az aks nodepool update --resource-group myResourceGroup --cluster myCluster --name pool1 --count 5

Solution

1. Step 1: Recall correct Azure CLI syntax for scaling node pools

   The correct command uses az aks nodepool scale with parameters for resource group, cluster name, node pool name, and node count.

2. Step 2: Match command with options

   az aks nodepool scale --resource-group myResourceGroup --cluster-name myCluster --name pool1 --node-count 5 matches the correct syntax exactly, including all required flags and parameters.

3. Final Answer:

   az aks nodepool scale --resource-group myResourceGroup --cluster-name myCluster --name pool1 --node-count 5 -> Option A

4. Quick Check:

   Correct CLI command = az aks nodepool scale --resource-group myResourceGroup --cluster-name myCluster --name pool1 --node-count 5 [OK]

Hint: Use 'az aks nodepool scale' with correct flags [OK]

Common Mistakes:

• Using wrong command verbs like 'update' or 'scale nodepool'
• Missing resource group or cluster name flags
• Incorrect flag names like --count instead of --node-count

3. Given this YAML snippet for an AKS node pool auto-scaling configuration:

nodePools:
  - name: pool1
    count: 3
    enableAutoScaling: true
    minCount: 2
    maxCount: 6

What is the maximum number of nodes this node pool can scale up to automatically?

medium

A. 6

B. 2

C. 3

D. Unlimited

Solution

1. Step 1: Identify auto-scaling parameters

   The YAML shows enableAutoScaling: true with minCount: 2 and maxCount: 6.

2. Step 2: Understand maxCount meaning

   The maxCount sets the upper limit for automatic scaling, so the node pool can scale up to 6 nodes.

3. Final Answer:

   6 -> Option A

4. Quick Check:

   Max nodes = maxCount = 6 [OK]

Hint: Max nodes = maxCount value in config [OK]

Common Mistakes:

• Confusing current count with maxCount
• Assuming unlimited scaling without maxCount
• Mixing minCount and maxCount values

4. You tried to scale a node pool using this command:
az aks nodepool scale --resource-group myRG --cluster-name myCluster --name poolA --node-count 0
But it failed. What is the most likely reason?

medium

A. Node pool name is case sensitive and must be uppercase

B. Missing --max-count parameter for scaling

C. Node count cannot be set to zero; minimum is 1

D. Cluster name is incorrect

Solution

1. Step 1: Understand node count limits

   AKS node pools require at least one node; zero nodes is invalid and causes failure.

2. Step 2: Check command parameters

   The command tries to scale to zero nodes, which is not allowed, causing the error.

3. Final Answer:

   Node count cannot be set to zero; minimum is 1 -> Option C

4. Quick Check:

   Minimum nodes = 1, zero not allowed [OK]

Hint: Node pools must have at least one node [OK]

Common Mistakes:

• Trying to scale node pool to zero nodes
• Assuming max-count is required for scaling
• Confusing case sensitivity of node pool names

5. You have an AKS cluster with two node pools: pool1 (3 nodes) and pool2 (5 nodes). You want to enable auto-scaling on pool2 to scale between 3 and 10 nodes based on workload. Which of the following steps correctly applies this configuration?

hard

A. Run az aks nodepool update --resource-group RG --cluster-name Cluster --name pool2 --enable-cluster-autoscaler --min-count 3 --max-count 10

B. Run az aks scale nodepool --resource-group RG --cluster Cluster --name pool2 --enable-auto-scaling --min 3 --max 10

C. Run az aks nodepool scale --resource-group RG --cluster-name Cluster --name pool2 --node-count 10

D. Run az aks nodepool update --resource-group RG --cluster-name Cluster --name pool2 --enable-auto-scaling --min-count 3 --max-count 10

Solution

1. Step 1: Identify correct command to enable auto-scaling

   The correct command is az aks nodepool update with flags --enable-auto-scaling, --min-count, and --max-count.

2. Step 2: Compare options for correct flag names and syntax

   Run az aks nodepool update --resource-group RG --cluster-name Cluster --name pool2 --enable-auto-scaling --min-count 3 --max-count 10 uses the correct flag --enable-auto-scaling (not cluster-autoscaler), and proper min/max count flags.

3. Final Answer:

   Run az aks nodepool update --resource-group RG --cluster-name Cluster --name pool2 --enable-auto-scaling --min-count 3 --max-count 10 -> Option D

4. Quick Check:

   Enable auto-scaling with correct flags = Run az aks nodepool update --resource-group RG --cluster-name Cluster --name pool2 --enable-auto-scaling --min-count 3 --max-count 10 [OK]

Hint: Use 'az aks nodepool update' with --enable-auto-scaling flag [OK]

Common Mistakes:

• Using wrong flag --enable-cluster-autoscaler
• Using scale command instead of update for auto-scaling
• Incorrect flag names like --min or --max instead of --min-count and --max-count