How to Deploy Qdrant Vector Database on Vultr Kubernetes Engine

Introduction

Qdrant is an open-source vector database designed to perform efficient similarity search operations on high-dimensional vector data. It allows users to store, search, and retrieve vector embeddings, which are numerical representations of text, images, audio, or other data types.

Qdrant Hybrid Cloud allows users to deploy Qdrant Clusters across user's self-hosted infrastructure, providing the feature of flexible deployment spanning multiple environments.

In this article, you are to create a Qdrant Hybrid Cloud Environment, deploy the Qdrant Cloud Agent in your Vultr Kubernetes Engine (VKE) Cluster. You will also be deploying a Qdrant Cluster along with installing the Nginx Ingress Controller and Cert Manager for requesting SSL certificate from Let's Encrypt in your VKE Cluster. At last, you will be applying custom API keys and scale your Qdrant Cluster.

Prerequisites

1. Deploy a Vultr Kubernetes Engine (VKE) Cluster.

2. Install Kubectl to access your VKE Cluster.

3. Install the Helm Package Manager.

Create a Qdrant Hybrid Cloud Environment

In this section, you are to create a Qdrant Hybrid Cloud Environment in your Qdrant dashboard, install Qdrant Cloud Agent in your VKE Cluster and deploy a Qdrant Cluster.

1. Select the Hybrid Cloud option from the left menu of your Qdrant dashboard.

2. Click on Get Started.

   

3. Enter a name for the Environment.

4. In the Kubernetes Configuration section, enter the name of the Kubernetes namespace for Qdrant components and leave other options on default.

5. Click on Create.

   

6. Once the Environment is created, generate single-use commands and copy them.

   

7. Execute these commands sequentially.

   • The first command creates a namespace in the VKE Cluster similar to the Hybrid Cloud Environment.
   • The next command deletes a Kubernetes secret qdrant-registry-creds if it already exists.
   • The third command also deletes a Kubernetes secret qdrant-cloud-creds if it already exists.
   • The fourth command creates a Docker Registry secret qdrant-registry-creds in the earlier created namespace.
   • The fifth command creates a generic secret qdrant-cloud-creds.
   • The second last command logs into a Helm Registry hosted at registry.cloud.qdrant.io with the credentials provided.
   • The last command installs a Helm chart named qdrant-cloud-agent from the specified path.

8. Once these commands are executed, refresh your Qdrant dashboard. You will be able to see the real-time state of the operator, agent, and Prometheus. Your Qdrant Hybrid Cloud Environment is ready to use when the state of the operator, agent, and Prometheus is changed to ready.

Create a Qdrant Cluster

Qdrant Cluster is a distributed deployment of the Qdrant Vector Database, in a cluster multiple Qdrant instances are coordinated together for scalability and high availability. The cluster has Qdrant nodes in it that are the individual instances of the Qdrant Database.

In this section, you are to create a Qdrant Cluster connected to your Hybrid Cloud Environment from your Qdrant dashboard.

1. Select Clusters option from the left menu of your Qdrant dashboard.

2. Click Create to create a Qdrant Cluster.

3. Select the Standard Plan, and Hybrid Cloud in the provider section.

4. Select the Hybrid Cloud Environment you created earlier as the region.

5. Enter a name for the cluster.

6. In the Cluster settings section, choose RAM per node as required. Choose the number of nodes, ideally the number of nodes in a Qdrant Cluster should be the same as the VKE Cluster. For the Disk space at least 45 GB is required.

7. Review the cluster configuration and Click Create.

8. The current status of the cluster can be observed from the Qdrant dashboard.

9. To follow the status of nodes on your VKE Cluster.

   CONSOLE

   Copy

   $ kubectl get nodes -n <qdrant-namespace>

   Explain Code

   Powered by Vultr AgentBeta

Create an Ingress Resource and Apply SSL

The Nginx Ingress Controller and Ingress resource in a Kubernetes cluster work together to manage external traffic to services inside a cluster. The Controller receives external traffic and routes it based on Ingress rules, and the Ingress resource defines the routing rules to route traffic to the backend services in the cluster.

The Cert Manager automates the management and issuance of TLS certificates from certificate authorities like Let's Encrypt.

In this section, you are to install the Nginx Ingress Controller and the Cert Manager and you are also going to apply the Kubernetes manifest to request SSL certificates from Let's Encrypt and the Ingress resource manifest with the routing rules to manage external traffic.

1. Add the Nginx Ingress repository.

   CONSOLE

   Copy

   $ helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx

   Explain Code

   Powered by Vultr AgentBeta

2. Install the Nginx Ingress Controller.

   CONSOLE

   Copy

   $ helm install ingress-nginx ingress-nginx/ingress-nginx

   Explain Code

   Powered by Vultr AgentBeta

3. After successful installation, a Vultr Load Balancer is automatically added to your cluster. Verify that the assigned external IP Address matches your Load Balancer IP in the VKE dashboard.

   CONSOLE

   Copy

   $ kubectl get services ingress-nginx-controller

   Explain Code

   Powered by Vultr AgentBeta

   Your output should look like the one below:

   OUTPUT

   Copy

   NAME                       TYPE           CLUSTER-IP       EXTERNAL-IP      PORT(S)                      AGE
   ingress-nginx-controller   LoadBalancer   10.108.201.180   192.0.2.1        80:31125/TCP,443:31287/TCP   39h

   Explain Code

   Powered by Vultr AgentBeta

   Copy the EXTERNAL-IP and configure your domain's A record to this IP.

4. Install the latest Cert Manager version.

   CONSOLE

   Copy

   $ kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.14.4/cert-manager.yaml

   Explain Code

   Powered by Vultr AgentBeta

5. Once installed, check the status of the pods created.

   CONSOLE

   Copy

   $ kubectl get pods --namespace cert-manager

   Explain Code

   Powered by Vultr AgentBeta

   Your output should look like the one below:

   OUTPUT

   Copy

   NAME                                       READY   STATUS    RESTARTS   AGE
   cert-manager-67c98b89c8-fd5w4              1/1     Running   0          4h12m
   cert-manager-cainjector-5c5695d979-xxffq   1/1     Running   0          4h12m
   cert-manager-webhook-7f9f8648b9-vsvwb      1/1     Running   0          4h12m

   Explain Code

   Powered by Vultr AgentBeta

6. Create a Let's Encrypt YAML manifest.

   CONSOLE

   Copy

   $ nano lets-encrypt.yaml

   Explain Code

   Powered by Vultr AgentBeta

7. Copy and paste the below contents.

   YAML

   Copy

   apiVersion: cert-manager.io/v1
   kind: ClusterIssuer
   metadata:
   name: letsencrypt-prod
   spec:
   acme:
       server: https://acme-v02.api.letsencrypt.org/directory
       email: <email-address>
       privateKeySecretRef:
       name: letsencrypt-prod
       solvers:
       - http01:
           ingress:
               class: nginx

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

   The above configuration creates a ClusterIssuer resource for issuing Let's Encrypt certificates. It uses the HTTP01 challenge solver to verify the ownership. Replace email-address with your actual email address.

8. Get the name of the cluster's StatefulSet.

   CONSOLE

   Copy

   $ kubectl get statefulset -n <qdrant-namespace>

   Explain Code

   Powered by Vultr AgentBeta

   Copy the name of the StatefulSet.

9. Create an Ingress Resource manifest.

   CONSOLE

   Copy

   $ nano qdrant-ingress.yaml

   Explain Code

   Powered by Vultr AgentBeta

10. Copy and paste the below contents.

    YAML

    Copy

    apiVersion: networking.k8s.io/v1
    kind: Ingress
    metadata:
    name: qdrant-ingress
    namespace: qdrant
    annotations:
        kubernetes.io/ingress.class: nginx
        cert-manager.io/cluster-issuer: "letsencrypt-prod"
    spec:
    ingressClassName: nginx
    tls:
    - hosts:
        - <domain-name>
        secretName: qdrant-ingress-tls
    rules:
    - host: <domain-name>
        http:
        paths:
        - path: /
            pathType: Prefix
            backend:
            service:
                name: <stateful-set>
                port:
                number: 6333   

    Explain Code

    Powered by Vultr AgentBeta

    Save and close the file.

    The Ingress resource allows external access to the Qdrant Cluster's StatefulSet resource. It uses the ClusterIssuer resource letsencrypt-prod to issue a new SSL certificate for domain-name and store it as a secret resource qdrant-ingress-tls. Replace domain-name with your actual domain name and stateful-set with the actual name retrieved in the earlier step.

11. Apply the issuer to the cluster.

    CONSOLE

    Copy

    $ kubectl apply -f lets-encrypt.yaml

    Explain Code

    Powered by Vultr AgentBeta

12. Verify the deployment.

    CONSOLE

    Copy

    $ kubectl get clusterissuer letsencrypt-prod

    Explain Code

    Powered by Vultr AgentBeta

    Your output should look like the one below:

    OUTPUT

    Copy

    NAME               READY   AGE
    letsencrypt-prod   True    4h13m

    Explain Code

    Powered by Vultr AgentBeta

13. Apply the Ingress resource.

    CONSOLE

    Copy

    $ kubectl apply -f qdrant-ingress.yaml

    Explain Code

    Powered by Vultr AgentBeta

14. Verify the Ingress resource.

    CONSOLE

    Copy

    $ kubectl get ingress -n <namespace>

    Explain Code

    Powered by Vultr AgentBeta

    Your output should look like the one below:

    OUTPUT

    Copy

    NAME                 CLASS   HOSTS                     ADDRESS           PORTS     AGE
    qdrant-ingress   nginx     <domain-name>     104.156.231.67   80, 443    4h16m

    Explain Code

    Powered by Vultr AgentBeta

    In a web browser, visit the domain URL and verify that you can access it over HTTPS.

    https://<domain-name>

Add Custom API Keys

In this section, you are to add custom API keys to access the Qdrant Databases. You are to create Kubernetes secret resources with the API key and also add them to your Qdrant Cluster configurations using the Qdrant dashboard.

1. Create a new Kubernetes secret.

   CONSOLE

   Copy

   $ nano qdrant-api-key.yaml

   Explain Code

   Powered by Vultr AgentBeta

2. Copy and paste the below contents.

   YAML

   Copy

   apiVersion: v1
   kind: Secret
   metadata:
       name: qdrant-api-keys
       namespace: <namespace>
   type: Opaque
   stringData:
       apikey: <api-key>

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

   The above configuration creates a Kubernetes secret resource to the API key that will be required to access and perform actions on the Qdrant Database. Replace api-key with an actual API key.

3. Create another Kubernetes secret.

   CONSOLE

   Copy

   $ nano qdrant-api-key-ro.yaml

   Explain Code

   Powered by Vultr AgentBeta

4. Copy and paste the below contents.

   YAML

   Copy

   apiVersion: v1
   kind: Secret
   metadata:
       name: qdrant-api-keys-ro
       namespace: <namespace>
   type: Opaque
   stringData:
       apikey: <api-key>

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

5. Apply the secret resources.

   CONSOLE

   Copy

   $ kubectl apply -f qdrant-api-key.yaml
   $ kubectl apply -f qdrant-api-key-ro.yaml

   Explain Code

   Powered by Vultr AgentBeta

6. Open the cluster details on your Qdrant dashboard by clicking on the cluster name.

7. Scroll down and select Configuration.

8. Enter qdrant-api-key as the secret name for the API Key and apikey as the secret key.

9. Next, Enter qdrant-api-key-ro as the secret name for the ReadOnly API Key and apikey as the secret key.

10. Click Save, after the newly added configurations are saved a rollout restart will be triggered for the Qdrant Cluster.

11. Once the cluster is healthy again, send a CURL request from your sessions terminal without the API key.

    CONSOLE

    Copy

    $ curl -X GET https://<domain-name>/collections

    Explain Code

    Powered by Vultr AgentBeta

    Your output should look like the one below:

    OUTPUT

    Copy

    Must provide an API key or an Authorization bearer

    Explain Code

    Powered by Vultr AgentBeta

12. Send a CURL request with the API key.

    CONSOLE

    Copy

    $ curl -X GET 'https://<domain-name>/collections' \
        --header 'api-key: <api-key>'

    Explain Code

    Powered by Vultr AgentBeta

    Your output should look like the one below if you have not made collections previously:

    OUTPUT

    Copy

    {"result":{"collections":[{"name":}]},"status":"ok","time":2.104e-6}

    Explain Code

    Powered by Vultr AgentBeta

    To learn more Qdrant Database operations

Scale the Qdrant Cluster

Scaling a cluster is important for maintaining performance, ensuring high availability and efficiently utilizing resources in production environments.

In this section, you are to scale your Qdrant Cluster.

1. In your Qdrant dashboard, open the Cluster Details.

2. Select Scale Up on the top right corner.

   

3. Scale up vertically by increasing the RAM per node and disk space. Vertically scaling a cluster results in a rolling restart as all nodes are affected by the new cluster configuration.

4. Or, scale up horizontally by increasing the number of nodes.

   

5. Review cluster configuration and click Scale up.

Conclusion

In this article, you created a Qdrant Hybrid Cloud Environment, deployed the Qdrant Cloud Agent in your VKE Cluster. You also deployed a Qdrant Cluster along with installation of the Nginx Ingress Controller and Cert Manager for requesting SSL certificate from Let's Encrypt in your VKE Cluster. At last, you applied custom API keys and scaled your Qdrant Cluster.