fix(collector): respect HPA scale; don't cap replicas at minReplicas

[shuheiktgw]

Description:
The issue was already well described in the original report. However, the operator currently always returns the values defined in spec, and autoscaler.minValues never scales up. I’ve fixed this problem. I also renamed the function to GetDesiredReplicas, since it’s not only called during initialization but also on every reconciliation.

Link to tracking Issue(s):

• Resolves: Autoscaler doesn't scale deployment when MinReplicas is set #4400

Testing:

Documentation:

[yuriolisa]

@shuheiktgw , thank you for this PR. Please ensure that we don't have a regression to this #2585 .

[yuriolisa]

I believe if GetDesiredReplicas will produce the same effect as described on #2585

[yuriolisa]

One idea would be adding a condition about the HPA here

[shuheiktgw]

Thank you so much for your comment! I tried reproducing the behavior described in the original issue, but from what I can see, even when we change minReplicas, the Operator doesn’t seem to create a new ReplicaSet, and the Pods aren’t restarted. Could you let me know if I’m missing something here?

apiVersion: opentelemetry.io/v1beta1
kind: OpenTelemetryCollector
metadata:
  name: simplest
spec:
  autoscaler:
    minReplicas: 2
    maxReplicas: 4
  config:
    receivers:
      otlp:
        protocols:
          grpc:
            endpoint: 0.0.0.0:4317
          http:
            endpoint: 0.0.0.0:4318
    processors:
      memory_limiter:
        check_interval: 1s
        limit_percentage: 75
        spike_limit_percentage: 15
      batch:
        send_batch_size: 10000
        timeout: 10s

    exporters:
      debug: {}

    service:
      pipelines:
        traces:
          receivers: [otlp]
          processors: [memory_limiter, batch]
          exporters: [debug]

NAME                                 READY   STATUS    RESTARTS   AGE
simplest-collector-c9fb74dcd-njqd8   1/1     Running   0          28s
simplest-collector-c9fb74dcd-plvd8   1/1     Running   0          28s

Update the minReplicas from 2 to 3.

spec:
  autoscaler:
    minReplicas: 3
    maxReplicas: 4

Add one Pod without restarting the others.

NAME                                 READY   STATUS    RESTARTS   AGE
simplest-collector-c9fb74dcd-6rt44   1/1     Running   0          6s
simplest-collector-c9fb74dcd-njqd8   1/1     Running   0          52s
simplest-collector-c9fb74dcd-plvd8   1/1     Running   0          52s

[schahal]

@yuriolisa / maintainers: do you think we can merge this change?

(This is the only fix I'm waiting on that's preventing me from upgrading the operator in my environment.)

Regarding the open review comment: I believe at worst @shuheiktgw's change simply reverts the bug that was introduced (and in my opinion that's more important than fixing the original issue the buggy PR was intended to fix because in reality minReplicas on HPAs aren't changing unless it's a conscious, one-off decision.