Correlating Traces with External IDs

Users can connect the traces logged on RagaAI Catalyst with their existing logs elsewhere - such as GCP, AWS, or other internal services

Enterprises usually maintain comprehensive logs in multiple places. For instance, workflow APIs could log the same LLM traces on an internal cloud service as well as on RagaAI Catalyst (if enabled) for observability. Using trace correlation as outlined below, users can now assign an external ID for each trace uploaded to Catalyst, for better maintainability and direct URL access.

query_list = [
    "Summarize the document in 10 words",
    "Summarize the document in 20 words",
    "Summarize the document in 30 words",
]
for i, external_id in enumerate(['exid_1', 'exid_2', 'exid_3']):
    tracer.set_external_id(external_id)
    main(query_list[i])

As shown, the set_external_id method can help assign new alphanumeric IDs programatically before calling your main function (assumed as a proxy for your LLM call). This will assign the last passed ID to the next trace that is uploaded.

Note that the ID should be updated before each call to avoid ambiguity. In the above example, the mapping would be something like:

External ID	Prompt
'exid_1'	"Summarize the document in 10 words"
'exid_2'	"Summarize the document in 20 words"
'exid_3'	"Summarize the document in 30 words"

Eventually, users can access the specific trace on RagaAI Catalyst directly from their central logs using a URL that would be something like:

https://catalyst.raga.ai/<project-name>/<dataset-name>/<external-id>/