Creating Direct Links for Specific Spans in Evaluation Logs

·Apr 11, 2025 02:18 PM

Hey there, simple question - Is there a way to construct a direct link for a specific span? I want to include it in my evaluation logs so it can easily be navigated to. I see that in the UI it's constructed kind of like this: {client.web_url}/projects/{some random project id that isn't actually the name of the project}/spans/{trace id in hex format}?selectedSpanNodeId={some span nodeid that isnt correlated with the actual span id} I see there's some code in here to get the project id but I don't see anything about getting the selectedSpanNodeId

5 comments

· Sorted by Oldest

RunLLM
APP
·
To construct a direct link for a specific span in Phoenix, you need both the project ID and the selectedSpanNodeId. The project ID can be retrieved using a GraphQL query, as shown in this example. However, the selectedSpanNodeId is not directly correlated with the actual span ID and is not explicitly mentioned in the data sources.Would you like more detail on how to retrieve the project ID or any other part of the process?Sources: - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/example/ExampleDetailsDialog.tsx - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/trace/TraceDetails.tsx
💡 Hint: Mention RunLLM in the thread for followups.
Jordan C.
·
ok I see here it is the result of a graphql query. I can look into doing that
🙌1
Jordan C.
·
I'm having trouble here due to the fact that the log_traces endpoint/otlp endpoint uses a background task to actually save the spans in the db, so they aren't immediately available. Is there a way to do this without needing to perform a retry operation? RunLLM
Jordan C.
·
RunLLM
RunLLM
APP
·
The issue you're facing with the log_traces endpoint using a background task means that spans are not immediately available for querying. Unfortunately, based on the available data sources, there is no direct method to bypass this delay without implementing a retry mechanism. The system is designed to process spans asynchronously to optimize performance and scalability.Would you like more detail on how to implement a retry mechanism or any other aspect of this process?Sources: - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/example/ExampleDetailsDialog.tsx - https://arize-ai.slack.com/archives/C04R3GXC8HK/p1724796536.922089
💡 Hint: Mention RunLLM in the thread for followups.

RunLLM
APP
·
To construct a direct link for a specific span in Phoenix, you need both the project ID and the selectedSpanNodeId. The project ID can be retrieved using a GraphQL query, as shown in this example. However, the selectedSpanNodeId is not directly correlated with the actual span ID and is not explicitly mentioned in the data sources.Would you like more detail on how to retrieve the project ID or any other part of the process?Sources: - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/example/ExampleDetailsDialog.tsx - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/trace/TraceDetails.tsx
💡 Hint: Mention RunLLM in the thread for followups.
Jordan C.
·
ok I see here it is the result of a graphql query. I can look into doing that
🙌1
Jordan C.
·
I'm having trouble here due to the fact that the log_traces endpoint/otlp endpoint uses a background task to actually save the spans in the db, so they aren't immediately available. Is there a way to do this without needing to perform a retry operation? RunLLM
Jordan C.
·
RunLLM
RunLLM
APP
·
The issue you're facing with the log_traces endpoint using a background task means that spans are not immediately available for querying. Unfortunately, based on the available data sources, there is no direct method to bypass this delay without implementing a retry mechanism. The system is designed to process spans asynchronously to optimize performance and scalability.Would you like more detail on how to implement a retry mechanism or any other aspect of this process?Sources: - https://github.com/Arize-ai/phoenix/blob/main/app/src/pages/example/ExampleDetailsDialog.tsx - https://arize-ai.slack.com/archives/C04R3GXC8HK/p1724796536.922089
💡 Hint: Mention RunLLM in the thread for followups.