fix(telemetry): export cache token metrics as strings #1000
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Problem
Cache token metrics are exported as objects instead of strings when zero. For example,
gen_ai.usage.cache_write_input_tokens
shows as{intValue: 0}
instead of"0"
, causing issues in trace analysis tools like Langfuse.Changes
Convert all token usage metrics to strings in tracer.py:
gen_ai.usage.prompt_tokens
gen_ai.usage.completion_tokens
gen_ai.usage.cache_read_input_tokens
gen_ai.usage.cache_write_input_tokens
Updated test assertions accordingly.
Testing
All 62 tracer tests pass. Verified that zero cache tokens now export as
"0"
(string) instead of objects.Fixes #971