Managing Multiple Projects in a Single Phoenix Instance

Ruchika B. · 2024-07-09T06:56:16.999Z

Hi, can we have multiple projects for a single phoenix instance. Like now I have projects/default can I have prjects/dev1/dev2/dev3 If yes then how can i acheive that

13 comments

· Sorted by Oldest

Nabeegh A.
·
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
Dustin N.
·
thanks Nabeegh! Alternatively we provide a using_project context manager that is convenient for working in a notebook environment, and you want different sells to send traces to different projects
Ruchika B.
·
Yes it worked for me thanks a lot for the help
Ruchika B.
·
One more thing actually I am trying to run some queries and below are my invocation params ss where streaming is not set to false Is this the reason for not getting the token count
Dustin N.
·
a lot of providers will not supply the token count when streaming
Ruchika B.
·
How can I get the token count then ?
Dustin N.
·
if you're using streaming and the provider doesn't supply that information at the moment you either have to estimate using a token counting tool or see if you can use the model without streaming
Dustin N.
·
on our roadmap is to provide utilities to track tokens even if the model provider does not supply all that info
Ruchika B.
·
Llamaindex supports token count ?
Dustin N.
·
it depends, I believe they support it for some models
Ruchika B.
·
I am using the gpt-35-turbo model
Dustin N.
·
I don't stay up to date on what modes llamaindex model wrappers support token counts for, since this depends on both OpenAI and llamaindex
Dustin N.
·
if llamaindex provides a token count in their callback, we will display it

Nabeegh A.
·
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
Dustin N.
·
thanks Nabeegh! Alternatively we provide a using_project context manager that is convenient for working in a notebook environment, and you want different sells to send traces to different projects
Ruchika B.
·
Yes it worked for me thanks a lot for the help
Ruchika B.
·
One more thing actually I am trying to run some queries and below are my invocation params ss where streaming is not set to false Is this the reason for not getting the token count
Dustin N.
·
a lot of providers will not supply the token count when streaming
Ruchika B.
·
How can I get the token count then ?
Dustin N.
·
if you're using streaming and the provider doesn't supply that information at the moment you either have to estimate using a token counting tool or see if you can use the model without streaming
Dustin N.
·
on our roadmap is to provide utilities to track tokens even if the model provider does not supply all that info
Ruchika B.
·
Llamaindex supports token count ?
Dustin N.
·
it depends, I believe they support it for some models
Ruchika B.
·
I am using the gpt-35-turbo model
Dustin N.
·
I don't stay up to date on what modes llamaindex model wrappers support token counts for, since this depends on both OpenAI and llamaindex
Dustin N.
·
if llamaindex provides a token count in their callback, we will display it