Keda-Operator : ScaledObject doesn't have correct scaleTargetRef specification although my specifications are correct

[kherath17]

A clear and concise description of what the bug is.
In a Kubernetes cluster I have created deployment each for Selenium-Hub and Chrome , and the respective services are also created and currently my tests are running on this setup. I wanted to auto-scale the pods depending on the session que requests and I used Keda - Selenium-Grid for that , but unfortunately Im running into an issue which says ScaledObject doesn't have correct scaleTargetRef specification , all the objects such as scaled objects , pods are on the same namespace

Expected Behavior

Scaled Object should create without Issue

Actual Behavior

Shows ScaledObject doesn't have correct scaleTargetRef specification by Keda-Operator

My .yaml files are as below

apiVersion: keda.sh/v1alpha1 kind: ScaledObject metadata: name: selenium-scaledobject namespace: default labels: deploymentName: selenium-chrome-node spec: scaleTargetRef: name: selenium-chrome-node minReplicaCount: 0 maxReplicaCount: 80 pollingInterval: 50 triggers: - type: selenium-grid metadata: url: 'http://selenium-hub-service:80/graphql' browserName: 'chrome'

PS - the url in the metadata section is the url of the hub service that I have created in my cluster and its respective port

Steps to Reproduce the Problem

1.Make 2 deployments in a Kubernetes cluster for Hub and the Node
2.Create the respective services to be access locally (Port Forwarding)
3.Setup the ScaledObject as described above pointing to the Chrome Node Deployment

Specifications

• KEDA Version: Keda 2.11.2
• Platform & Version: Windows 10 X64
• Kubernetes Version: v1.24.16-eks-2d98532
• Scaler(s): Keda - ScaledObject

[JorTurFer]

I move this issue to KEDA repo because it's not related with the helm chart

[JorTurFer]

Could you share KEDA operator logs? It looks like the deployment doesn't exist, but I think that we will see a more clear message in KEDA operator logs

[kherath17]

@JorTurFer I have missed to add the name: selenium-node-chrome which is mentioned under scaletargetref spec in my chrome deployment labels, this is working now

[kumarnaveentulasi]

Hello can u help me i was facing the same issue the issue was as follows

Type Reason Age From Message

---

Warning ScaledObjectCheckFailed 47m (x9 over 47m) keda-operator Target resource doesn't exist
Warning ScaledObjectCheckFailed 47m (x9 over 47m) keda-operator ScaledObject doesn't have correct scaleTargetRef specification
Normal KEDAScalersStarted 47m keda-operator Started scalers watch
Normal ScaledObjectReady 47m (x3 over 47m) keda-operator ScaledObject is ready for scaling
Warning KEDAScalerFailed 47m keda-operator error inspecting rabbitMQ: Exception (404) Reason: "NOT_FOUND - no queue 'inference_queue' in vhost '/'"
Normal KEDAScalersStarted 2m31s (x719 over 47m) keda-operator Scaler rabbitmq is built.

but my hpa was as follows

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
name: my-microservice-scaledobject
namespace: dev
spec:
scaleTargetRef:
name: my-microservice # The name of your deployment
pollingInterval: 5 # Time interval to poll RabbitMQ queue for messages
cooldownPeriod: 30 # Time to wait before scaling down after scaling up
minReplicaCount: 0 # Minimum replicas can scale to 0
maxReplicaCount: 20 # Maximum replicas (adjust based on your needs)
triggers:

• type: rabbitmq
  metadata:
  host: "amqp://admin:Truefan@[email protected]:5672" # Correct RabbitMQ URL
  queueName: "inference_queue" # The RabbitMQ queue name to monitor
  vhost: "naveen-test-dev" # The vhost name (ensure this is the correct vhost in RabbitMQ)
  queueLength: "1" # Scale 1 pod per message in the queue

i was the file name as hpa.yml and the deploying in the dev namespace and keda operator in the keda namespace can any help me with this