New to KubeDB? Please start here.
Autoscaling the Compute Resource of a Kafka Combined Cluster
This guide will show you how to use KubeDB to autoscale compute resources i.e. cpu and memory of a Kafka combined cluster.
Before You Begin
At first, you need to have a Kubernetes cluster, and the
kubectlcommand-line tool must be configured to communicate with your cluster.Install
KubeDBProvisioner, Ops-manager and Autoscaler operator in your cluster following the steps here.Install
Metrics Serverfrom hereYou should be familiar with the following
KubeDBconcepts:
To keep everything isolated, we are going to use a separate namespace called demo throughout this tutorial.
$ kubectl create ns demo
namespace/demo created
Note: YAML files used in this tutorial are stored in docs/examples/kafka directory of kubedb/docs repository.
Autoscaling of Combined Cluster
Here, we are going to deploy a Kafka Combined Cluster using a supported version by KubeDB operator. Then we are going to apply KafkaAutoscaler to set up autoscaling.
Deploy Kafka Combined Cluster
In this section, we are going to deploy a Kafka Topology database with version 3.6.1. Then, in the next section we will set up autoscaling for this database using KafkaAutoscaler CRD. Below is the YAML of the Kafka CR that we are going to create,
apiVersion: kubedb.com/v1
kind: Kafka
metadata:
name: kafka-dev
namespace: demo
spec:
replicas: 2
version: 3.6.1
podTemplate:
spec:
containers:
- name: kafka
resources:
limits:
memory: 1Gi
requests:
cpu: 500m
memory: 1Gi
storage:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 1Gi
storageClassName: standard
storageType: Durable
deletionPolicy: WipeOut
Let’s create the Kafka CRO we have shown above,
$ kubectl create -f https://github.com/kubedb/docs/raw/v2024.8.21/docs/examples/kafka/autoscaler/kafka-combined.yaml
kafka.kubedb.com/kafka-dev created
Now, wait until kafka-dev has status Ready. i.e,
$ kubectl get kf -n demo -w
NAME TYPE VERSION STATUS AGE
kafka-dev kubedb.com/v1 3.6.1 Provisioning 0s
kafka-dev kubedb.com/v1 3.6.1 Provisioning 24s
.
.
kafka-dev kubedb.com/v1 3.6.1 Ready 92s
Let’s check the Pod containers resources,
$ kubectl get pod -n demo kafka-dev-0 -o json | jq '.spec.containers[].resources'
{
"limits": {
"memory": "1Gi"
},
"requests": {
"cpu": "500m",
"memory": "1Gi"
}
}
Let’s check the Kafka resources,
$ kubectl get kafka -n demo kafka-dev -o json | jq '.spec.podTemplate.spec.containers[].resources'
{
"limits": {
"memory": "1Gi"
},
"requests": {
"cpu": "500m",
"memory": "1Gi"
}
}
You can see from the above outputs that the resources are same as the one we have assigned while deploying the kafka.
We are now ready to apply the KafkaAutoscaler CRO to set up autoscaling for this database.
Compute Resource Autoscaling
Here, we are going to set up compute resource autoscaling using a KafkaAutoscaler Object.
Create KafkaAutoscaler Object
In order to set up compute resource autoscaling for this combined cluster, we have to create a KafkaAutoscaler CRO with our desired configuration. Below is the YAML of the KafkaAutoscaler object that we are going to create,
apiVersion: autoscaling.kubedb.com/v1alpha1
kind: KafkaAutoscaler
metadata:
name: kf-combined-autoscaler
namespace: demo
spec:
databaseRef:
name: kafka-dev
opsRequestOptions:
timeout: 5m
apply: IfReady
compute:
node:
trigger: "On"
podLifeTimeThreshold: 5m
resourceDiffPercentage: 20
minAllowed:
cpu: 600m
memory: 1.5Gi
maxAllowed:
cpu: 1
memory: 2Gi
controlledResources: ["cpu", "memory"]
containerControlledValues: "RequestsAndLimits"
Here,
spec.databaseRef.namespecifies that we are performing compute resource scaling operation onkafka-devcluster.spec.compute.node.triggerspecifies that compute autoscaling is enabled for this cluster.spec.compute.node.podLifeTimeThresholdspecifies the minimum lifetime for at least one of the pod to initiate a vertical scaling.spec.compute.node.resourceDiffPercentagespecifies the minimum resource difference in percentage. The default is 10%. If the difference between current & recommended resource is less than ResourceDiffPercentage, Autoscaler Operator will ignore the updating.spec.compute.node.minAllowedspecifies the minimum allowed resources for the cluster.spec.compute.node.maxAllowedspecifies the maximum allowed resources for the cluster.spec.compute.node.controlledResourcesspecifies the resources that are controlled by the autoscaler.spec.compute.node.containerControlledValuesspecifies which resource values should be controlled. The default is “RequestsAndLimits”.spec.opsRequestOptionscontains the options to pass to the created OpsRequest. It has 2 fields.timeoutspecifies the timeout for the OpsRequest.applyspecifies when the OpsRequest should be applied. The default is “IfReady”.
Let’s create the KafkaAutoscaler CR we have shown above,
$ kubectl apply -f https://github.com/kubedb/docs/raw/v2024.8.21/docs/examples/kafka/autoscaler/compute/kafka-combined-autoscaler.yaml
kafkaautoscaler.autoscaling.kubedb.com/kf-combined-autoscaler created
Verify Autoscaling is set up successfully
Let’s check that the kafkaautoscaler resource is created successfully,
$ kubectl describe kafkaautoscaler kf-combined-autoscaler -n demo
Name: kf-combined-autoscaler
Namespace: demo
Labels: <none>
Annotations: <none>
API Version: autoscaling.kubedb.com/v1alpha1
Kind: KafkaAutoscaler
Metadata:
Creation Timestamp: 2024-08-27T05:55:51Z
Generation: 1
Owner References:
API Version: kubedb.com/v1
Block Owner Deletion: true
Controller: true
Kind: Kafka
Name: kafka-dev
UID: a0153c7f-1e1e-4070-a318-c7c1153b810a
Resource Version: 1104655
UID: 817602cc-f851-4fc5-b2c1-1d191462ac56
Spec:
Compute:
Node:
Container Controlled Values: RequestsAndLimits
Controlled Resources:
cpu
memory
Max Allowed:
Cpu: 1
Memory: 2Gi
Min Allowed:
Cpu: 600m
Memory: 1536Mi
Pod Life Time Threshold: 5m0s
Resource Diff Percentage: 20
Trigger: On
Database Ref:
Name: kafka-dev
Ops Request Options:
Apply: IfReady
Timeout: 5m0s
Status:
Checkpoints:
Cpu Histogram:
Bucket Weights:
Index: 0
Weight: 4610
Index: 1
Weight: 10000
Reference Timestamp: 2024-08-27T05:55:00Z
Total Weight: 0.35081120875606336
First Sample Start: 2024-08-27T05:55:44Z
Last Sample Start: 2024-08-27T05:56:49Z
Last Update Time: 2024-08-27T05:57:10Z
Memory Histogram:
Reference Timestamp: 2024-08-27T06:00:00Z
Ref:
Container Name: kafka
Vpa Object Name: kafka-dev
Total Samples Count: 3
Version: v3
Conditions:
Last Transition Time: 2024-08-27T05:56:32Z
Message: Successfully created kafkaOpsRequest demo/kfops-kafka-dev-z8d3l5
Observed Generation: 1
Reason: CreateOpsRequest
Status: True
Type: CreateOpsRequest
Vpas:
Conditions:
Last Transition Time: 2024-08-27T05:56:10Z
Status: True
Type: RecommendationProvided
Recommendation:
Container Recommendations:
Container Name: kafka
Lower Bound:
Cpu: 600m
Memory: 1536Mi
Target:
Cpu: 600m
Memory: 1536Mi
Uncapped Target:
Cpu: 100m
Memory: 511772986
Upper Bound:
Cpu: 1
Memory: 2Gi
Vpa Name: kafka-dev
Events: <none>
So, the kafkaautoscaler resource is created successfully.
you can see in the Status.VPAs.Recommendation section, that recommendation has been generated for our database. Our autoscaler operator continuously watches the recommendation generated and creates an kafkaopsrequest based on the recommendations, if the database pods resources are needed to scaled up or down.
Let’s watch the kafkaopsrequest in the demo namespace to see if any kafkaopsrequest object is created. After some time you’ll see that a kafkaopsrequest will be created based on the recommendation.
$ watch kubectl get kafkaopsrequest -n demo
Every 2.0s: kubectl get kafkaopsrequest -n demo
NAME TYPE STATUS AGE
kfops-kafka-dev-z8d3l5 VerticalScaling Progressing 10s
Let’s wait for the ops request to become successful.
$ kubectl get kafkaopsrequest -n demo
NAME TYPE STATUS AGE
kfops-kafka-dev-z8d3l5 VerticalScaling Successful 3m2s
We can see from the above output that the KafkaOpsRequest has succeeded. If we describe the KafkaOpsRequest we will get an overview of the steps that were followed to scale the cluster.
$ kubectl describe kafkaopsrequests -n demo kfops-kafka-dev-z8d3l5
Name: kfops-kafka-dev-z8d3l5
Namespace: demo
Labels: app.kubernetes.io/component=database
app.kubernetes.io/instance=kafka-dev
app.kubernetes.io/managed-by=kubedb.com
app.kubernetes.io/name=kafkas.kubedb.com
Annotations: <none>
API Version: ops.kubedb.com/v1alpha1
Kind: KafkaOpsRequest
Metadata:
Creation Timestamp: 2024-08-27T05:56:32Z
Generation: 1
Owner References:
API Version: autoscaling.kubedb.com/v1alpha1
Block Owner Deletion: true
Controller: true
Kind: KafkaAutoscaler
Name: kf-combined-autoscaler
UID: 817602cc-f851-4fc5-b2c1-1d191462ac56
Resource Version: 1104871
UID: 8b7615c6-d38b-4d5a-b733-6aa93cd41a29
Spec:
Apply: IfReady
Database Ref:
Name: kafka-dev
Timeout: 5m0s
Type: VerticalScaling
Vertical Scaling:
Node:
Resources:
Limits:
Memory: 1536Mi
Requests:
Cpu: 600m
Memory: 1536Mi
Status:
Conditions:
Last Transition Time: 2024-08-27T05:56:32Z
Message: Kafka ops-request has started to vertically scaling the kafka nodes
Observed Generation: 1
Reason: VerticalScaling
Status: True
Type: VerticalScaling
Last Transition Time: 2024-08-27T05:56:35Z
Message: Successfully updated PetSets Resources
Observed Generation: 1
Reason: UpdatePetSets
Status: True
Type: UpdatePetSets
Last Transition Time: 2024-08-27T05:56:40Z
Message: get pod; ConditionStatus:True; PodName:kafka-dev-0
Observed Generation: 1
Status: True
Type: GetPod--kafka-dev-0
Last Transition Time: 2024-08-27T05:56:40Z
Message: evict pod; ConditionStatus:True; PodName:kafka-dev-0
Observed Generation: 1
Status: True
Type: EvictPod--kafka-dev-0
Last Transition Time: 2024-08-27T05:57:10Z
Message: check pod running; ConditionStatus:True; PodName:kafka-dev-0
Observed Generation: 1
Status: True
Type: CheckPodRunning--kafka-dev-0
Last Transition Time: 2024-08-27T05:57:15Z
Message: get pod; ConditionStatus:True; PodName:kafka-dev-1
Observed Generation: 1
Status: True
Type: GetPod--kafka-dev-1
Last Transition Time: 2024-08-27T05:57:16Z
Message: evict pod; ConditionStatus:True; PodName:kafka-dev-1
Observed Generation: 1
Status: True
Type: EvictPod--kafka-dev-1
Last Transition Time: 2024-08-27T05:57:25Z
Message: check pod running; ConditionStatus:True; PodName:kafka-dev-1
Observed Generation: 1
Status: True
Type: CheckPodRunning--kafka-dev-1
Last Transition Time: 2024-08-27T05:57:30Z
Message: Successfully Restarted Pods With Resources
Observed Generation: 1
Reason: RestartPods
Status: True
Type: RestartPods
Last Transition Time: 2024-08-27T05:57:30Z
Message: Successfully completed the vertical scaling for kafka
Observed Generation: 1
Reason: Successful
Status: True
Type: Successful
Observed Generation: 1
Phase: Successful
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Starting 4m33s KubeDB Ops-manager Operator Start processing for KafkaOpsRequest: demo/kfops-kafka-dev-z8d3l5
Normal Starting 4m33s KubeDB Ops-manager Operator Pausing Kafka databse: demo/kafka-dev
Normal Successful 4m33s KubeDB Ops-manager Operator Successfully paused Kafka database: demo/kafka-dev for KafkaOpsRequest: kfops-kafka-dev-z8d3l5
Normal UpdatePetSets 4m30s KubeDB Ops-manager Operator Successfully updated PetSets Resources
Warning get pod; ConditionStatus:True; PodName:kafka-dev-0 4m25s KubeDB Ops-manager Operator get pod; ConditionStatus:True; PodName:kafka-dev-0
Warning evict pod; ConditionStatus:True; PodName:kafka-dev-0 4m25s KubeDB Ops-manager Operator evict pod; ConditionStatus:True; PodName:kafka-dev-0
Warning check pod running; ConditionStatus:False; PodName:kafka-dev-0 4m19s KubeDB Ops-manager Operator check pod running; ConditionStatus:False; PodName:kafka-dev-0
Warning check pod running; ConditionStatus:True; PodName:kafka-dev-0 3m55s KubeDB Ops-manager Operator check pod running; ConditionStatus:True; PodName:kafka-dev-0
Warning get pod; ConditionStatus:True; PodName:kafka-dev-1 3m50s KubeDB Ops-manager Operator get pod; ConditionStatus:True; PodName:kafka-dev-1
Warning evict pod; ConditionStatus:True; PodName:kafka-dev-1 3m49s KubeDB Ops-manager Operator evict pod; ConditionStatus:True; PodName:kafka-dev-1
Warning check pod running; ConditionStatus:False; PodName:kafka-dev-1 3m45s KubeDB Ops-manager Operator check pod running; ConditionStatus:False; PodName:kafka-dev-1
Warning check pod running; ConditionStatus:True; PodName:kafka-dev-1 3m40s KubeDB Ops-manager Operator check pod running; ConditionStatus:True; PodName:kafka-dev-1
Normal RestartPods 3m35s KubeDB Ops-manager Operator Successfully Restarted Pods With Resources
Normal Starting 3m35s KubeDB Ops-manager Operator Resuming Kafka database: demo/kafka-dev
Normal Successful 3m35s KubeDB Ops-manager Operator Successfully resumed Kafka database: demo/kafka-dev for KafkaOpsRequest: kfops-kafka-dev-z8d3l5
Now, we are going to verify from the Pod, and the Kafka yaml whether the resources of the topology database has updated to meet up the desired state, Let’s check,
$ kubectl get pod -n demo kafka-dev-0 -o json | jq '.spec.containers[].resources'
{
"limits": {
"memory": "1536Mi"
},
"requests": {
"cpu": "600m",
"memory": "1536Mi"
}
}
$ kubectl get kafka -n demo kafka-dev -o json | jq '.spec.podTemplate.spec.containers[].resources'
{
"limits": {
"memory": "1536Mi"
},
"requests": {
"cpu": "600m",
"memory": "1536Mi"
}
}
The above output verifies that we have successfully auto scaled the resources of the Kafka combined cluster.
Cleaning Up
To clean up the Kubernetes resources created by this tutorial, run:
kubectl delete kafkaopsrequest -n demo kfops-kafka-dev-z8d3l5
kubectl delete kafkaautoscaler -n demo kf-combined-autoscaler
kubectl delete kf -n demo kafka-dev
kubectl delete ns demo
Next Steps
- Detail concepts of Kafka object.
- Different Kafka topology clustering modes here.
- Monitor your Kafka database with KubeDB using out-of-the-box Prometheus operator.
- Want to hack on KubeDB? Check our contribution guidelines.