New to KubeDB? Please start here.

Autoscaling the Compute Resource of a MongoDB Sharded Database

This guide will show you how to use KubeDB to autoscale compute resources i.e. cpu and memory of a MongoDB sharded database.

Before You Begin

  • At first, you need to have a Kubernetes cluster, and the kubectl command-line tool must be configured to communicate with your cluster.

  • Install KubeDB Community, Enterprise and Autoscaler operator in your cluster following the steps here.

  • Install Metrics Server from here

  • Install Vertical Pod Autoscaler from here

  • You should be familiar with the following KubeDB concepts:

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/mongodb directory of kubedb/docs repository.

Autoscaling of Sharded Database

Here, we are going to deploy a MongoDB sharded database using a supported version by KubeDB operator. Then we are going to apply MongoDBAutoscaler to set up autoscaling.

Deploy MongoDB Sharded Database

In this section, we are going to deploy a MongoDB sharded database with version 4.2.3. Then, in the next section we will set up autoscaling for this database using MongoDBAutoscaler CRD. Below is the YAML of the MongoDB CR that we are going to create,

apiVersion: kubedb.com/v1alpha2
kind: MongoDB
metadata:
  name: mg-sh
  namespace: demo
spec:
  version: "4.2.3"
  storageType: Durable
  shardTopology:
    configServer:
      storage:
        resources:
          requests:
            storage: 1Gi
      replicas: 3
      podTemplate:
        spec:
          resources:
            requests:
              cpu: "200m"
              memory: "300Mi"
    mongos:
      replicas: 2
      podTemplate:
        spec:
          resources:
            requests:
              cpu: "200m"
              memory: "300Mi"
    shard:
      storage:
        resources:
          requests:
            storage: 1Gi
      replicas: 3
      shards: 2
      podTemplate:
        spec:
          resources:
            requests:
              cpu: "200m"
              memory: "300Mi"
  terminationPolicy: WipeOut

Let’s create the MongoDB CRO we have shown above,

$ kubectl create -f https://github.com/kubedb/docs/raw/v2021.11.24/docs/examples/mongodb/autoscaling/compute/mg-sh.yaml
mongodb.kubedb.com/mg-sh created

Now, wait until mg-sh has status Ready. i.e,

$ kubectl get mg -n demo
NAME    VERSION    STATUS    AGE
mg-sh   4.2.3      Ready     3m57s

Let’s check a shard Pod containers resources,

$ kubectl get pod -n demo mg-sh-shard0-0 -o json | jq '.spec.containers[].resources'
{
  "limits": {
    "cpu": "200m",
    "memory": "300Mi"
  },
  "requests": {
    "cpu": "200m",
    "memory": "300Mi"
  }
}

Let’s check the MongoDB resources,

$ kubectl get mongodb -n demo mg-sh -o json | jq '.spec.shardTopology.shard.podTemplate.spec.resources'
{
  "limits": {
    "cpu": "200m",
    "memory": "300Mi"
  },
  "requests": {
    "cpu": "200m",
    "memory": "300Mi"
  }
}

You can see from the above outputs that the resources are same as the one we have assigned while deploying the mongodb.

We are now ready to apply the MongoDBAutoscaler CRO to set up autoscaling for this database.

Compute Resource Autoscaling

Here, we are going to set up compute resource autoscaling using a MongoDBAutoscaler Object.

Create MongoDBAutoscaler Object

In order to set up compute resource autoscaling for the shard pod of the database, we have to create a MongoDBAutoscaler CRO with our desired configuration. Below is the YAML of the MongoDBAutoscaler object that we are going to create,

apiVersion: autoscaling.kubedb.com/v1alpha1
kind: MongoDBAutoscaler
metadata:
  name: mg-as-sh
  namespace: demo
spec:
  databaseRef:
    name: mg-sh
  compute:
    shard:
      trigger: "On"
      podLifeTimeThreshold: 5m
      minAllowed:
        cpu: 250m
        memory: 350Mi
      maxAllowed:
        cpu: 1
        memory: 1Gi
      controlledResources: ["cpu", "memory"]

Here,

  • spec.databaseRef.name specifies that we are performing compute resource scaling operation on mg-sh database.
  • spec.compute.shard.trigger specifies that compute autoscaling is enabled for the shard pods of this database.
  • spec.compute.shard.podLifeTimeThreshold specifies the minimum lifetime for at least one of the pod to initiate a vertical scaling.
  • spec.compute.shard.minAllowed specifies the minimum allowed resources for the database.
  • spec.compute.shard.maxAllowed specifies the maximum allowed resources for the database.
  • spec.compute.shard.controlledResources specifies the resources that are controlled by the autoscaler.

Note: In this demo we are only setting up the autoscaling for the shard pods, that’s why we only specified the shard section of the autoscaler. You can enable autoscaling for mongos and configServer pods in the same yaml, by specifying the spec.mongos and spec.configServer section, similar to the spec.shard section we have configured in this demo.

Let’s create the MongoDBAutoscaler CR we have shown above,

$ kubectl apply -f https://github.com/kubedb/docs/raw/v2021.11.24/docs/examples/mongodb/autoscaling/compute/mg-as-sh.yaml
mongodbautoscaler.autoscaling.kubedb.com/mg-as-sh created

Verify Autoscaling is set up successfully

Let’s check that the mongodbautoscaler resource is created successfully,

$ kubectl get mongodbautoscaler -n demo
NAME        AGE
mg-as-sh    102s

$ kubectl describe mongodbautoscaler mg-as-sh -n demo
Name:         mg-as-sh
Namespace:    demo
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.kubedb.com/v1alpha1
Kind:         MongoDBAutoscaler
Metadata:
  Creation Timestamp:  2021-03-07T16:49:09Z
  Generation:          1
  Managed Fields:
    API Version:  autoscaling.kubedb.com/v1alpha1
    Fields Type:  FieldsV1
    fieldsV1:
      f:metadata:
        f:annotations:
          .:
          f:kubectl.kubernetes.io/last-applied-configuration:
      f:spec:
        .:
        f:compute:
          .:
          f:shard:
            .:
            f:controlledResources:
            f:maxAllowed:
              .:
              f:cpu:
              f:memory:
            f:minAllowed:
              .:
              f:cpu:
              f:memory:
            f:podLifeTimeThreshold:
            f:trigger:
        f:databaseRef:
          .:
          f:name:
    Manager:      kubectl-client-side-apply
    Operation:    Update
    Time:         2021-03-07T16:49:09Z
    API Version:  autoscaling.kubedb.com/v1alpha1
    Fields Type:  FieldsV1
    fieldsV1:
      f:status:
        .:
        f:conditions:
    Manager:         kubedb-autoscaler
    Operation:       Update
    Time:            2021-03-07T16:50:13Z
  Resource Version:  879550
  Self Link:         /apis/autoscaling.kubedb.com/v1alpha1/namespaces/demo/mongodbautoscalers/mg-as-sh
  UID:               7e6880f1-42ba-4d78-ba1c-02aa9ea522e9
Spec:
  Compute:
    Shard:
      Controlled Resources:
        cpu
        memory
      Max Allowed:
        Cpu:     1
        Memory:  1Gi
      Min Allowed:
        Cpu:                    250m
        Memory:                 350Mi
      Pod Life Time Threshold:  5m0s
      Trigger:                  On
  Database Ref:
    Name:  mg-sh
Status:
Events:                    <none>

So, the mongodbautoscaler resource is created successfully.

Now, lets verify that the vertical pod autoscaler (vpa) resource is created successfully,

$ kubectl get vpa -n demo
NAME                AGE
vpa-mg-sh-shard0    110s
vpa-mg-sh-shard1    110s

$ kubectl describe vpa vpa-mg-sh-shard0  -n demo
Name:         vpa-mg-sh-shard0
Namespace:    demo
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2021-03-07T16:49:09Z
  Generation:          2
  Managed Fields:
    API Version:  autoscaling.k8s.io/v1
    Fields Type:  FieldsV1
    fieldsV1:
      f:metadata:
        f:ownerReferences:
          .:
          k:{"uid":"7e6880f1-42ba-4d78-ba1c-02aa9ea522e9"}:
            .:
            f:apiVersion:
            f:blockOwnerDeletion:
            f:controller:
            f:kind:
            f:name:
            f:uid:
      f:spec:
        .:
        f:resourcePolicy:
          .:
          f:containerPolicies:
        f:targetRef:
          .:
          f:apiVersion:
          f:kind:
          f:name:
        f:updatePolicy:
          .:
          f:updateMode:
      f:status:
    Manager:      kubedb-autoscaler
    Operation:    Update
    Time:         2021-03-07T16:49:09Z
    API Version:  autoscaling.k8s.io/v1
    Fields Type:  FieldsV1
    fieldsV1:
      f:status:
        f:conditions:
        f:recommendation:
          .:
          f:containerRecommendations:
    Manager:    recommender
    Operation:  Update
    Time:       2021-03-07T16:50:03Z
  Owner References:
    API Version:           autoscaling.kubedb.com/v1alpha1
    Block Owner Deletion:  true
    Controller:            true
    Kind:                  MongoDBAutoscaler
    Name:                  mg-as-sh
    UID:                   7e6880f1-42ba-4d78-ba1c-02aa9ea522e9
  Resource Version:        879512
  Self Link:               /apis/autoscaling.k8s.io/v1/namespaces/demo/verticalpodautoscalers/vpa-mg-sh-shard0
  UID:                     e73e9920-5c4d-4e8e-887e-38b06120c9a6
Spec:
  Resource Policy:
    Container Policies:
      Container Name:  mongodb
      Controlled Resources:
        cpu
        memory
      Controlled Values:  RequestsAndLimits
      Max Allowed:
        Cpu:     1
        Memory:  1Gi
      Min Allowed:
        Cpu:     250m
        Memory:  350Mi
  Target Ref:
    API Version:  apps/v1
    Kind:         StatefulSet
    Name:         mg-sh-shard0
  Update Policy:
    Update Mode:  Off
Status:
  Conditions:
    Last Transition Time:  2021-03-07T16:50:03Z
    Status:                False
    Type:                  RecommendationProvided
  Recommendation:
Events:          <none>

So, we can verify from the above output that two vpa resources are created for our two shard successfully, but you can see that the RecommendationProvided condition is false and also the Recommendation section of the vpa is empty. Let’s wait some time and describe the vpa again.

$ kubectl describe vpa vpa-mg-sh-shard0  -n demo
Name:         vpa-mg-sh-shard0
Namespace:    demo
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2021-03-07T16:49:09Z
  Generation:          2
  Managed Fields:
    API Version:  autoscaling.k8s.io/v1
    Fields Type:  FieldsV1
    fieldsV1:
      f:metadata:
        f:ownerReferences:
          .:
          k:{"uid":"7e6880f1-42ba-4d78-ba1c-02aa9ea522e9"}:
            .:
            f:apiVersion:
            f:blockOwnerDeletion:
            f:controller:
            f:kind:
            f:name:
            f:uid:
      f:spec:
        .:
        f:resourcePolicy:
          .:
          f:containerPolicies:
        f:targetRef:
          .:
          f:apiVersion:
          f:kind:
          f:name:
        f:updatePolicy:
          .:
          f:updateMode:
      f:status:
    Manager:      kubedb-autoscaler
    Operation:    Update
    Time:         2021-03-07T16:49:09Z
    API Version:  autoscaling.k8s.io/v1
    Fields Type:  FieldsV1
    fieldsV1:
      f:status:
        f:conditions:
        f:recommendation:
          .:
          f:containerRecommendations:
    Manager:    recommender
    Operation:  Update
    Time:       2021-03-07T16:50:03Z
  Owner References:
    API Version:           autoscaling.kubedb.com/v1alpha1
    Block Owner Deletion:  true
    Controller:            true
    Kind:                  MongoDBAutoscaler
    Name:                  mg-as-sh
    UID:                   7e6880f1-42ba-4d78-ba1c-02aa9ea522e9
  Resource Version:        879512
  Self Link:               /apis/autoscaling.k8s.io/v1/namespaces/demo/verticalpodautoscalers/vpa-mg-sh-shard0
  UID:                     e73e9920-5c4d-4e8e-887e-38b06120c9a6
Spec:
  Resource Policy:
    Container Policies:
      Container Name:  mongodb
      Controlled Resources:
        cpu
        memory
      Controlled Values:  RequestsAndLimits
      Max Allowed:
        Cpu:     1
        Memory:  1Gi
      Min Allowed:
        Cpu:     250m
        Memory:  350Mi
  Target Ref:
    API Version:  apps/v1
    Kind:         StatefulSet
    Name:         mg-sh-shard0
  Update Policy:
    Update Mode:  Off
Status:
  Conditions:
    Last Transition Time:  2021-03-07T16:50:03Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  mongodb
      Lower Bound:
        Cpu:     250m
        Memory:  350Mi
      Target:
        Cpu:     250m
        Memory:  350Mi
      Uncapped Target:
        Cpu:     203m
        Memory:  262144k
      Upper Bound:
        Cpu:     1
        Memory:  1Gi
Events:          <none>

As you can see from the output the vpa has generated a recommendation for the shard pod of the database. Our autoscaler operator continuously watches the recommendation generated and creates an mongodbopsrequest based on the recommendations, if the database pods are needed to scaled up or down.

Let’s watch the mongodbopsrequest in the demo namespace to see if any mongodbopsrequest object is created. After some time you’ll see that a mongodbopsrequest will be created based on the recommendation.

$ watch kubectl get mongodbopsrequest -n demo
Every 2.0s: kubectl get mongodbopsrequest -n demo
NAME                          TYPE              STATUS       AGE
mops-vpa-mg-sh-shard-3uqbrq   VerticalScaling   Progressing  19s

Let’s wait for the ops request to become successful.

$ watch kubectl get mongodbopsrequest -n demo
Every 2.0s: kubectl get mongodbopsrequest -n demo
NAME                            TYPE              STATUS       AGE
mops-vpa-mg-sh-shard-3uqbrq     VerticalScaling   Successful   5m8s

We can see from the above output that the MongoDBOpsRequest has succeeded. If we describe the MongoDBOpsRequest we will get an overview of the steps that were followed to scale the database.

$ kubectl describe mongodbopsrequest -n demo mops-vpa-mg-sh-shard-3uqbrq
Name:         mops-vpa-mg-sh-shard-3uqbrq
Namespace:    demo
Labels:       app.kubernetes.io/component=database
              app.kubernetes.io/instance=mg-sh
              app.kubernetes.io/managed-by=kubedb.com
              app.kubernetes.io/name=mongodbs.kubedb.com
Annotations:  <none>
API Version:  ops.kubedb.com/v1alpha1
Kind:         MongoDBOpsRequest
Metadata:
  Creation Timestamp:  2021-03-07T16:50:13Z
  Generation:          1
  Managed Fields:
    API Version:  ops.kubedb.com/v1alpha1
    Fields Type:  FieldsV1
    fieldsV1:
      f:metadata:
        f:labels:
          .:
          f:app.kubernetes.io/component:
          f:app.kubernetes.io/instance:
          f:app.kubernetes.io/managed-by:
          f:app.kubernetes.io/name:
        f:ownerReferences:
      f:spec:
        .:
        f:configuration:
        f:databaseRef:
          .:
          f:name:
        f:type:
        f:verticalScaling:
          .:
          f:shard:
            .:
            f:limits:
              .:
              f:cpu:
              f:memory:
            f:requests:
              .:
              f:cpu:
              f:memory:
    Manager:      kubedb-autoscaler
    Operation:    Update
    Time:         2021-03-07T16:50:13Z
    API Version:  ops.kubedb.com/v1alpha1
    Fields Type:  FieldsV1
    fieldsV1:
      f:status:
        .:
        f:conditions:
        f:observedGeneration:
        f:phase:
    Manager:    kubedb-enterprise
    Operation:  Update
    Time:       2021-03-07T16:50:13Z
  Owner References:
    API Version:           autoscaling.kubedb.com/v1alpha1
    Block Owner Deletion:  true
    Controller:            true
    Kind:                  MongoDBAutoscaler
    Name:                  mg-as-sh
    UID:                   7e6880f1-42ba-4d78-ba1c-02aa9ea522e9
  Resource Version:        880864
  Self Link:               /apis/ops.kubedb.com/v1alpha1/namespaces/demo/mongodbopsrequests/mops-vpa-mg-sh-shard-3uqbrq
  UID:                     a9eb9a92-3a93-441c-90b9-a272cfff4e85
Spec:
  Configuration:
  Database Ref:
    Name:  mg-sh
  Type:    VerticalScaling
  Vertical Scaling:
    Shard:
      Limits:
        Cpu:     250m
        Memory:  350Mi
      Requests:
        Cpu:     250m
        Memory:  350Mi
Status:
  Conditions:
    Last Transition Time:  2021-03-07T16:50:13Z
    Message:               MongoDB ops request is vertically scaling database
    Observed Generation:   1
    Reason:                VerticalScaling
    Status:                True
    Type:                  VerticalScaling
    Last Transition Time:  2021-03-07T16:50:13Z
    Message:               Successfully updated StatefulSets Resources
    Observed Generation:   1
    Reason:                UpdateStatefulSetResources
    Status:                True
    Type:                  UpdateStatefulSetResources
    Last Transition Time:  2021-03-07T16:55:21Z
    Message:               Successfully Vertically Scaled Shard Resources
    Observed Generation:   1
    Reason:                UpdateShardResources
    Status:                True
    Type:                  UpdateShardResources
    Last Transition Time:  2021-03-07T16:55:21Z
    Message:               Successfully Vertically Scaled Database
    Observed Generation:   1
    Reason:                Successful
    Status:                True
    Type:                  Successful
  Observed Generation:     1
  Phase:                   Successful
Events:
  Type    Reason                      Age    From                        Message
  ----    ------                      ----   ----                        -------
  Normal  PauseDatabase               14m    KubeDB Enterprise Operator  Pausing MongoDB demo/mg-sh
  Normal  PauseDatabase               14m    KubeDB Enterprise Operator  Successfully paused MongoDB demo/mg-sh
  Normal  Starting                    14m    KubeDB Enterprise Operator  Updating Resources of StatefulSet: mg-sh-shard0
  Normal  Starting                    14m    KubeDB Enterprise Operator  Updating Resources of StatefulSet: mg-sh-shard1
  Normal  UpdateStatefulSetResources  14m    KubeDB Enterprise Operator  Successfully updated StatefulSets Resources
  Normal  UpdateShardResources        9m13s  KubeDB Enterprise Operator  Successfully Vertically Scaled Shard Resources
  Normal  ResumeDatabase              9m13s  KubeDB Enterprise Operator  Resuming MongoDB demo/mg-sh
  Normal  ResumeDatabase              9m13s  KubeDB Enterprise Operator  Successfully resumed MongoDB demo/mg-sh
  Normal  Successful                  9m13s  KubeDB Enterprise Operator  Successfully Vertically Scaled Database

Now, we are going to verify from the Pod, and the MongoDB yaml whether the resources of the shard pod of the database has updated to meet up the desired state, Let’s check,

$ kubectl get pod -n demo mg-sh-shard0-0 -o json | jq '.spec.containers[].resources'
{
  "limits": {
    "cpu": "250m",
    "memory": "350Mi"
  },
  "requests": {
    "cpu": "250m",
    "memory": "350Mi"
  }
}

$ kubectl get mongodb -n demo mg-sh -o json | jq '.spec.shardTopology.shard.podTemplate.spec.resources'
{
  "limits": {
    "cpu": "250m",
    "memory": "350Mi"
  },
  "requests": {
    "cpu": "250m",
    "memory": "350Mi"
  }
}

The above output verifies that we have successfully auto scaled the resources of the MongoDB sharded database.

Cleaning Up

To clean up the Kubernetes resources created by this tutorial, run:

kubectl delete mg -n demo mg-sh
kubectl delete mongodbautoscaler -n demo mg-as-sh