Helm charts for K8s deployment of Infinity

[indranilr]

Although dockerized deployments are supported, an official helm chart would be very handy for K8s deployments. Is this already available ?

[greptile-apps]

Currently, there are no Helm charts available for deploying Infinity on Kubernetes. To create a Helm chart for Infinity, follow these steps:

1. Create Directory Structure:

   mkdir -p helm/infinity/templates
   touch helm/infinity/Chart.yaml helm/infinity/values.yaml helm/infinity/templates/deployment.yaml helm/infinity/templates/service.yaml

2. Define Chart Metadata (helm/infinity/Chart.yaml):

   apiVersion: v2
   name: infinity
   description: A Helm chart for Kubernetes deployment of Infinity
   version: 0.1.0

3. Set Default Values (helm/infinity/values.yaml):

   image:
     repository: michaelf34/infinity
     tag: latest
   service:
     type: ClusterIP
     port: 80

4. Create Deployment Template (helm/infinity/templates/deployment.yaml):

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: infinity
   spec:
     replicas: 1
     selector:
       matchLabels:
         app: infinity
     template:
       metadata:
         labels:
           app: infinity
       spec:
         containers:
         - name: infinity
           image: {{ .Values.image.repository }}:{{ .Values.image.tag }}
           ports:
           - containerPort: {{ .Values.service.port }}
           env:
           - name: MODEL_ID
             value: "BAAI/bge-small-en-v1.5"
           - name: PORT
             value: "80"

5. Create Service Template (helm/infinity/templates/service.yaml):

   apiVersion: v1
   kind: Service
   metadata:
     name: infinity
   spec:
     type: {{ .Values.service.type }}
     ports:
     - port: {{ .Values.service.port }}
       targetPort: 80
     selector:
       app: infinity

6. Package and Deploy:

   helm package helm/infinity
   helm install infinity ./infinity-0.1.0.tgz

References

/.github/ISSUE_TEMPLATE/bug-report.yml
/.github/ISSUE_TEMPLATE
/docs/docs/deploy.md
/.github/ISSUE_TEMPLATE/new-model-addition.yml
/.github/workflows/release.yaml

/README.md
/docs/docs
/docs/assets/openapi.json
/docs/docs/index.md
/docs
/libs/infinity_emb/tests/data/benchmark
/docs/assets
/.github/ISSUE_TEMPLATE/feature-request.yml
/libs/infinity_emb
/libs/infinity_emb/Makefile
/libs/infinity_emb/README.md
/docs/benchmarks/simple_app.py
/libs/infinity_emb/tests/end_to_end/test_authentication.py
/docs/benchmarks
/libs/infinity_emb/tests/end_to_end/test_optimum_embedding.py
/libs
/libs/infinity_emb/tests/data
/.github/workflows/ci.yaml
/.github

About Greptile

This response provides a starting point for your research, not a precise solution.

Help us improve! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

Ask Greptile · Edit Issue Bot Settings

[michaelfeil]

I agree, but have not used helm recently. There is no helm chart available, potentially the infra folder could be a suitable place?

[samos123]

I am working on finishing the KubeAI integration that allows you to deploy Infinity easily on K8s: https://github.com/substratusai/kubeai

PR that's soon to be merged: substratusai/kubeai#197