There are changes in the Terraform plan / apply when the autoscaler is currently scaling up.

[mickaeltr]

Hello,

We have applications with the autoscaler enabled:

resource "scalingo_container_type" "agir_frontend" {
  app    = scalingo_app.agir_frontend.id
  name   = "web"
  amount = 2
  size   = "S"
}

resource "scalingo_autoscaler" "agir_frontend" {
  app            = scalingo_app.agir_frontend.id
  container_type = scalingo_container_type.agir_frontend.name
  min_containers = scalingo_container_type.agir_frontend.amount
  max_containers = 10
  metric         = "memory"
  target         = 0.9
  disabled       = false
}

I noticed that, when the autoscaler is currently scaling up (e.g. 4 containers), there is a change in Terraform plan / apply:

scalingo_container_type = amount : 4 → 2

I also noticed that when doing a second plan / apply soon after, there is another change (that is more concerning…):

scalingo_autoscaler = disabled : true → false

I should point out that during this entire time, we did not make any changes to the Terraform configuration or Scalingo settings. This is simply the automatic behavior of the autoscaler.

I don't know what the behavior should ideally be, but it seems surprising to me that changes appear and that it takes 2 plan / apply cycles to restore stability.

Thank you