CODEWITHSUNDEEP
azureazure-virtual-machineazure-vm-scale-setazure-autoscaling-block
Potis23
Potis23

Reputation: 512

Azure Autoscaling Duration Time and Cool Down Time

I keep reading contradictory oppinions on how autoscaling works in Azure with regards to duration time and cooldown time.

Till now I have concluded that if the metric threshold is met after the cool down time then a new instance is immediately added (or removed), but Im might be wrong so please correct me.

Example:

| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 |
| 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 |
| 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 |
| 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 |
| 51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 |

Duration = 10 minutes
Cool down = 5 minutes
Scale out threshold = 70% CPU usage
Starting with 1 instance (assumed)
1 instance is added in every scale out

Lets say that, since we started a machine the CPU usage is constatly > 70% for 60 mins:

Im sure that:

1 instance is added at the end of the duration time which is 10 mins
So, min|11| = 2 instances

Before a new instance will be added we have to wait for the 5 mins of cooldown time.
Azure keeps collecting metrics during the cooldown period.

This is where the confusion starts:

Scenario1:

1 instance is added at the end of the cooldown time which is 5 mins
(Since the criteria is met of this period | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 |)
So, min|16| = 3 instances(
min|11|(end of duration0) 2 instances -> 
min|16|(end of duration0 + cooldown0) 3 instances
)

So, an instance will be added every 5 mins (taking into account that the CPU usage is constantly over 70%)

Next one will be added at min|22|, min|28|,....

Scenario2:

1 instance is added at the end of the cooldown time which is 5 mins + 10 mins of duration time which is 10 mins
So, min|26| = 3 instances(
min|11|(end of duration0) 2 instances -> 
min|16|(end of duration0 + cooldown0) 2 instances ->
min|26|(end of duration0 + end of cooldown0 + end of duration01) 3 instances
)

So, an instance will be added every 15 mins (taking into account that the CPU usage is constantly over 70%)

Next one will be added at min|42|, min|58|,....

Please only reply if you are sure, there is enough confusion out there already. Real example with screenshots would be helpful

Upvotes: 1

Views: 279

Answers (2)

Az-fta
Az-fta

Reputation: 1

Scenario 1 is the correct one. Try it with a scaleset. Add the rule for CPU>5%, period 5mins, cool down 2mins and you will notice that every 2 mins you will be getting a new instance.

Upvotes: 0

Joel Odey
Joel Odey

Reputation: 92

The correct behavior aligns with Scenario 2. After the initial scaling action, the system waits for the cooldown period to end before starting a new duration evaluation. Therefore, an instance is added every 15 minutes if the CPU usage consistently exceeds 70%.

For more information:

-https://learn.microsoft.com/en-us/azure/azure-monitor/autoscale/autoscale-understanding-settings

-https://learn.microsoft.com/en-us/azure/azure-monitor/autoscale/autoscale-best-practices

If above links are not of any help I would suggest raising a support request directly to Azure for further assistance as sometimes scaling can be tricky in terms of behaviour.

Upvotes: 1

Related Questions