feat: add custom restful backend for complex scenarios (e.g, rag)

[popsiclexu]

📑 Description

Add a new backend that allows access to a custom API, enabling the integration of personalized services into k8sgpt. For example, it supports the integration of RAG services to enhance the accuracy of responses based local knowledge.

✅ Checks

• My pull request adheres to the code style of this project
• My code requires changes to the documentation
• I have updated the documentation as required
• All the tests have passed

ℹ Additional Information

Usage

# ./bin/k8sgpt auth add --backend customrest --baseurl http://localhost:8090/completion --model llama3.1
customrest added to the AI backend provider list
# ./k8sgpt analyze --backend customrest --explain --filter=Pod
100% |█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| (2/2, 8 it/min)
AI Provider: customrest

0: Pod cluster-sample-ns/minio-harborcluster-sample-zone-harbor-0()
- Error: 0/4 nodes are available: 4 pod has unbound immediate PersistentVolumeClaims.
Error: No available nodes to schedule pods, unbound PersistentVolumeClaims exist.

Solution: 
1. Check PersistentVolumeClaims (PVCs) and delete any unused ones.
2. Update PVCs with correct storage class or request.
3. Scale down or delete pods using the problematic PVCs.

You cant get a tutorial for customrest backend from https://github.com/popsiclexu/k8sgpt-docs/blob/feat/customrest/docs/tutorials/custom-rest-backend.md.
And the docs pr is k8sgpt-ai/docs#120

[matthisholleville]

Thank you for your PR—great work! We need to discuss further how we'll support RAG as an AI Provider.

For example, what types of authentication will we support? How should we include tests and tutorials in our documentation? etc.

By the way, your PR is an excellent starting point for this discussion.

[popsiclexu]

Thank you for your PR—great work! We need to discuss further how we'll support RAG as an AI Provider.

For example, what types of authentication will we support? How should we include tests and tutorials in our documentation? etc.

By the way, your PR is an excellent starting point for this discussion.

I added a tutorial for this: https://github.com/popsiclexu/k8sgpt-docs/blob/feat/customrest/docs/tutorials/custom-rest-backend.md.
And the docs pr is k8sgpt-ai/docs#120

[arbreezy]

@popsiclexu is this not possible with localai as a backend ?

[popsiclexu]

@arbreezy The custom backend allows users to implement more complex and customizable functionality. Additionally, it enables integration with any vector database or embedding model within their own code, offering greater flexibility and extensibility.

[AlexsJones]

I am happy to see this PR, and willing to accept it, but I think we need an OpenAPI or similar schema to help others to understand our thinking here of how the protocol might look. Or at the very least some example documentation. Would you be willing to put something together @popsiclexu ?

[popsiclexu]

Thank you for your suggestion! I’ve added a description of the OpenAPI schema to the tutorial documentation. k8sgpt-ai/docs#135 @AlexsJones