Nabeegh A.

Kratos - Since that requires you changing the Phoenix code, you will first have to fork the Phoenix repository from github.com/arize-ai/phoenix. Once you do, you can explore the Development.md file on instruction of how to setup the project locally. Once that is done, you can then go to changing the code. Afterwards, you will have to create a new build of the system. First in the phoenix/app, you can create a build using pnpm run build Then in the main folder, you can use hatch build to build the whole system. You can then build a new docker image, or install the built python package in your application.

At the moment no, the Phoenix team has it on their roadmap that they want to use Vite, but as of now, you will have to create phoenix builds every time you want to see the new changes.

Kratos - You can refer to this page in the documentation: https://docs.arize.com/phoenix/deployment/deploying-phoenix It has all the details. You can setup Phoenix in docker using:

docker pull arizephoenix/phoenix
docker run -p 6006:6006 -p 4317:4317 -i -t arizephoenix/phoenix:latest

Then send traces using OpenInference as described in the documentation page. You will be able to see them on the Phoenix dashboard.

If you are talking about creating multiple projects, you can simply do that by using this in your openinference code:

resource = Resource(attributes={ResourceAttributes.PROJECT_NAME: project_name})
tracer_provider = trace_sdk.TracerProvider(resource=resource)

Where project_name would be your project name. In this way, if you were to send some spans using a project name "dev", you would see two projects in the dashboard now, default and dev. If you are referring to creating projects inside a project, I am not sure if phoenix supports that at the moment. You can still achieve this behavior by using a prefix name in your projects. Ref: https://docs.arize.com/phoenix/deployment/deploying-phoenix#initialize-instrumentation

Sounds good Roméo I.. Could you share what workaround you used? Always interesting to see some clever tricks

Hi Roméo I. - Could you clarify what exactly do you mean by "not stall". Do you want the exporter to wait until the service is reachable? Or not cause an error? With that said, I think this issue is not necessarily a phoenix issue. You can use some tools from the programming language you are using to tackle this.

Assuming you were able to get the deployment running according to guidelines on Phoenix deployment page, you can just go to the URL in your browser. The URL depends on how you configured kubernetes and what ports you used. The default address for phoenix dashboard is http://localhost:6006/

Can you please let me know if you are doing a local docker deployment or using a service?

Could you clarify what are you trying to do? Are you looking to access the phoenix dashboard in the browser? Or export traces from your application to it? In any case, did you follow the guide on https://docs.arize.com/phoenix/deployment/deploying-phoenix? This has all the info to properly setup phoenix in a container and export traces to it. If you are running locally, you can go to localhost:6006 in your browser (provided you forward port 6006 in your container)

Since the data generated and inputs given to the system are very diverse, you cannot use a fixed set of rules to evaluate the system. For evaluating RAG pipelines you need to use LLM-Evals. Phoenix already has a Retrieval (RAG) Relevance evaluation template available that you can use. You can also check out RAGAS.