App Engine Flex auto-scaling help

[Nigel Grange]

Hi,

I'm trying to understand how the metrics on the App Engine dashboard relate to the autoscaling properties set in the app.yaml.

The latest deployed version has the following properties:

automatic_scaling:

  cool_down_period: 120s

  min_num_instances: 1

  max_num_instances: 3

  cpu_utilization:

    target_utilization: 0.6

network: {}

resources:

  cpu: 1

  memory_gb: 4

  disk_size_gb: 10

However, in the dashboard, I'm seeing a sustained 4 instances, peaking at around 7 - graph is attached.

My question is - should I be expecting the dashboard to be showing a maximum of 3 instances (I'd expect so), and if not, how do I get the autoscaling to work as expected?

Many Thanks!

[George (Cloud Platform Support)]

Flexible environment uses the Compute Engine Autoscaler. Flexible environment does not support all of the autoscaling options that are available to Compute Engine. The autoscaler continuously collects usage information based on the policy, compares actual utilization to your desired target utilization, and determines if the group needs to be scaled up or down. If you set target_utilization: 0.6, and this usage parameter is exceeded, new instances will get spun up.