hey @Marc
hey @Marc
yesterday my team had a zoom meeting with you, where they asked about the missing traces,
let me give you more context so you can help us better.
i have made a simple RAG app using langchain, where langfuse is not able to trace output, PFA screensnap
5 Replies
by investigating further on traces, where we got output traced and where not, i found that the trace about "StrOutputParser" is missing in the trace where outout is not traced. PFA screenshot of both the traces.
Super helpful. This is most likely and issue with flushing the data to the Langfuse api. Have a look here for more details: https://langfuse.com/faq/all/missing-traces
Thanks, it works.
but need help in another issue.
yesterday i have added only 4 metrics in my evaluation configs, today i have added some more by customising, problem is when i ran a sample rap code it is able to evaluate on new metrics also which i have added today but my rag application only evaluate using 4 metrics which i have added yesterday. why my rag application is not able to capture new metrics?? PFA screenshots.
and sometime it is not able to evaluate, it only take 10 sec to show evaluation, but not consistent in my scenario?
would love to have a look together in case you have 5 minutes to demo this issue: cal.com/marc-kl/5