Node pools and auto scaling in GCP - Step-by-Step Execution

Process Flow - Node pools and auto scaling

Create Node Pool

↓

Set Min and Max Nodes

↓

Monitor Cluster Load

↓

Is Load High?

No→Keep Current Nodes

Yes↓

Scale Up Nodes

↓

Is Load Low?

No→Keep Current Nodes

Yes↓

Scale Down Nodes

↓

Repeat Monitoring

This flow shows how a node pool is created with min and max nodes, then the system monitors load and scales nodes up or down automatically.

Execution Sample

GCP

gcloud container node-pools create pool-1 --cluster=my-cluster --num-nodes=3 --enable-autoscaling --min-nodes=1 --max-nodes=5

# Cluster load increases
# Autoscaler adds nodes

# Cluster load decreases
# Autoscaler removes nodes

This code creates a node pool with autoscaling enabled, then nodes are added or removed based on cluster load.

Process Table

Step	Cluster Load	Current Nodes	Autoscaler Decision	Action Taken
1	Low	3	No scaling needed	Nodes remain 3
2	High	3	Scale up	Nodes increased to 4
3	High	4	Scale up	Nodes increased to 5
4	High	5	Max nodes reached	Nodes remain 5
5	Medium	5	No scaling needed	Nodes remain 5
6	Low	5	Scale down	Nodes decreased to 4
7	Low	4	Scale down	Nodes decreased to 3
8	Low	3	Scale down	Nodes decreased to 2
9	Low	2	Scale down	Nodes decreased to 1
10	Low	1	Min nodes reached	Nodes remain 1

💡 Autoscaler stops scaling when min or max node limits are reached

Status Tracker

Variable	Start	After 1	After 2	After 3	After 4	After 5	After 6	After 7	After 8	After 9	Final
Cluster Load	Low	High	High	High	Medium	Low	Low	Low	Low	Low	Low
Current Nodes	3	3	4	5	5	5	4	3	2	1	1
Autoscaler Decision	No scaling	Scale up	Scale up	Max nodes	No scaling	Scale down	Scale down	Scale down	Scale down	Min nodes	Min nodes

Key Moments - 3 Insights

Why does the autoscaler stop adding nodes at step 4 even though the load is high?

At step 4, the autoscaler reached the maximum node limit (5 nodes), so it cannot add more nodes despite high load, as shown in the execution_table row 4.

Why does the autoscaler not reduce nodes below 1 at step 10?

The autoscaler respects the minimum node limit (1 node), so it stops scaling down at step 10, as seen in execution_table row 10.

What happens if the cluster load stays medium at step 5?

When load is medium, autoscaler decides no scaling is needed, so nodes remain the same, as shown in execution_table row 5.

Visual Quiz - 3 Questions

Look at the execution_table, how many nodes are running after step 3?

A 3 nodes

B 4 nodes

C 5 nodes

D 6 nodes

💡 Hint

Check the 'Current Nodes' column at step 3 in the execution_table.

At which step does the autoscaler first scale down nodes?

A Step 6

B Step 4

C Step 5

D Step 7

💡 Hint

Look for the first 'Scale down' decision in the 'Autoscaler Decision' column.

If the minimum nodes were set to 2 instead of 1, what would be the node count at step 10?

A 1 node

B 2 nodes

C 3 nodes

D 0 nodes

💡 Hint

Refer to the 'Min nodes reached' logic in the execution_table and variable_tracker.

Concept Snapshot

Node pools group worker machines in a cluster.
Autoscaling adjusts node count based on load.
Set minimum and maximum nodes to control scaling.
Autoscaler adds nodes when load is high.
Autoscaler removes nodes when load is low.
Scaling stops at min or max node limits.

Full Transcript

Node pools are groups of worker machines in a cloud cluster. You create a node pool and set minimum and maximum nodes to control autoscaling. The system watches the cluster load. If load is high, it adds nodes up to the maximum limit. If load is low, it removes nodes down to the minimum limit. This process repeats to keep resources efficient. Autoscaling stops adding nodes when max nodes are reached and stops removing nodes when min nodes are reached.