Setting lagThreshold to 0 causes the keda-operator to panic

[Stii]

Report

The lagThreshold for a scaledobject was set to 0 in an attempt to disable the SO, but this caused the keda-operator to panic, resulting in the operator going into a CrashloopBackoff state.

Expected Behavior

Warn that the scaledobject has an incorrect value or prevent lagThreshold to be set to 0

Actual Behavior

The keda-operator produces panic: runtime error: integer divide by zero and goes into a CrashloopBackoff state

Steps to Reproduce the Problem

1. Edit a scaledobject
2. Set the lagThreshold to 0

Logs from KEDA operator

panic: runtime error: integer divide by zero

goroutine 1021 [running]:
github.com/kedacore/keda/v2/pkg/scalers.(*kafkaScaler).GetMetrics(0xc0079837a0, {0x1, 0x4}, {0x3435e92, 0xb}, {0x0, 0xa})
        /workspace/pkg/scalers/kafka_scaler.go:485 +0x4ac
github.com/kedacore/keda/v2/pkg/scaling/cache.(*ScalersCache).getScaledJobMetrics(0xc00384a550, {0x3ae9008, 0xc0079ed680}, 0xc0078d0a00)
        /workspace/pkg/scaling/cache/scalers_cache.go:274 +0x523
github.com/kedacore/keda/v2/pkg/scaling/cache.(*ScalersCache).IsScaledJobActive(0xc00074f8b0, {0x3ae9008, 0xc0079ed680}, 0xc0078d0a00)
        /workspace/pkg/scaling/cache/scalers_cache.go:124 +0xda
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).checkScalers(0xc000617340, {0x3ae9008, 0xc0079ed680}, {0x3375f00, 0xc0078d0a00}, {0x3ac7030, 0xc0079f3810})
        /workspace/pkg/scaling/scale_handler.go:286 +0x317
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).startScaleLoop(0xc000617340, {0x3ae9008, 0xc0079ed680}, 0xc0078d8640, {0x3375f00, 0xc0078d0a00}, {0x3ac7030, 0xc0079f3810})
        /workspace/pkg/scaling/scale_handler.go:149 +0x31c
created by github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).HandleScalableObject
        /workspace/pkg/scaling/scale_handler.go:108 +0x465

KEDA Version

2.7.1

Kubernetes Version

1.19

Platform

Other

Scaler Details

Kafka

Anything else?

No response

[zroubalik]

Thanks for reporting, though I am not able to reproduce the problem, if I set lagThreshold to 0, the HPA is not even created:

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: test
spec:
  minReplicaCount: 0
  maxReplicaCount: 5
  scaleTargetRef:
    name: test
  triggers:
  - type: kafka
    metadata:
      topic: my-topic
      bootstrapServers: my-cluster-kafka.svc:9092
      consumerGroup: my-group
      lagThreshold: '0'
      offsetResetPolicy: 'latest'

ERROR   controller.scaledobject Failed to update HPA    {"reconciler group": "keda.sh", "reconciler kind": "ScaledObject", "name": "test", "namespace": "test", "HPA.Namespace": "test", "HPA.Name": "keda-hpa-test", "error": "HorizontalPodAutoscaler.autoscaling \"keda-hpa-test\" is invalid: spec.metrics[0].external.target.averageValue: Invalid value: resource.Quantity{i:resource.int64Amount{value:0, scale:0}, d:resource.infDecAmount{Dec:(*inf.Dec)(nil)}, s:\"0\", Format:\"DecimalSI\"}: must be positive"}
keda-operator-78b964d4cd-z9zvd keda-operator github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).ensureHPAForScaledObjectExists
keda-operator-78b964d4cd-z9zvd keda-operator    /workspace/controllers/keda/scaledobject_controller.go:385
keda-operator-78b964d4cd-z9zvd keda-operator github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).reconcileScaledObject
keda-operator-78b964d4cd-z9zvd keda-operator    /workspace/controllers/keda/scaledobject_controller.go:231
keda-operator-78b964d4cd-z9zvd keda-operator github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).Reconcile
keda-operator-78b964d4cd-z9zvd keda-operator    /workspace/controllers/keda/scaledobject_controller.go:182
keda-operator-78b964d4cd-z9zvd keda-operator sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile
keda-operator-78b964d4cd-z9zvd keda-operator    /go/pkg/mod/sigs.k8s.io/controller-runtime@v0.11.2/pkg/internal/controller/controller.go:114
keda-operator-78b964d4cd-z9zvd keda-operator sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler
keda-operator-78b964d4cd-z9zvd keda-operator    /go/pkg/mod/sigs.k8s.io/controller-runtime@v0.11.2/pkg/internal/controller/controller.go:311
keda-operator-78b964d4cd-z9zvd keda-operator sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem
keda-operator-78b964d4cd-z9zvd keda-operator    /go/pkg/mod/sigs.k8s.io/controller-runtime@v0.11.2/pkg/internal/controller/controller.go:266
keda-operator-78b964d4cd-z9zvd keda-operator sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2
keda-operator-78b964d4cd-z9zvd keda-operator    /go/pkg/mod/sigs.k8s.io/controller-runtime@v0.11.2/pkg/internal/controller/controller.go:227

But adding an extra check before the division doesn't hurt.

[zroubalik]

@Stii also, to pause scaling, you should use the annotation:
https://keda.sh/docs/2.7/concepts/scaling-deployments/#pause-autoscaling

[Stii]

Ah, thank you so much!