[opentelemetry-instrumentation-genai-anthropic] Record cache token usage via the shared invocation fields - #641
Merged
lmolkova merged 2 commits intoSep 9, 2026
Conversation
Krishnachaitanyakc
force-pushed
the
anthropic-cache-tokens-via-invocation-fields
branch
from
September 6, 2026 23:37
7806cc9 to
441736a
Compare
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
It removes previously importable module-level constants in messages_extractors.py, which is potentially a breaking change for downstream imports unless kept as backward-compatible aliases.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Pull request overview
Updates the Anthropic GenAI instrumentation to record cache token usage via the shared InferenceInvocation fields (instead of package-local attribute-name string constants), aligning emission with semantic conventions while keeping emitted attribute names the same.
Changes:
- Switch
messages_extractors.set_invocation_response_attributes()to populateInferenceInvocation.cache_creation_input_tokens/cache_read_input_tokens. - Remove the locally defined cache-usage attribute name constants from
messages_extractors.py. - Update sync/async messages tests to assert using semconv
GenAIAttributesconstants; add a changelog fragment.
File summaries
| File | Description |
|---|---|
| instrumentation/opentelemetry-instrumentation-genai-anthropic/tests/test_sync_messages.py | Updates cache token assertions to use semconv GenAIAttributes constants. |
| instrumentation/opentelemetry-instrumentation-genai-anthropic/tests/test_async_messages.py | Updates cache token assertions to use semconv GenAIAttributes constants and removes old constant imports. |
| instrumentation/opentelemetry-instrumentation-genai-anthropic/src/opentelemetry/instrumentation/genai/anthropic/messages_extractors.py | Sets cache token usage via shared invocation fields; removes local attribute-name constants. |
| instrumentation/opentelemetry-instrumentation-genai-anthropic/.changelog/641.fixed | Documents the behavior/implementation change. |
Review details
- Files reviewed: 4/4 changed files
- Comments generated: 1
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Pull request dashboard statusWaiting on reviewers · refreshed 2026-09-07 01:32 UTC Review the latest changes. Status above doesn't look right?
|
lmolkova
approved these changes
Sep 8, 2026
…age via the shared invocation fields Set cache_creation_input_tokens / cache_read_input_tokens on the invocation instead of writing locally defined attribute-name constants into invocation.attributes, matching how the openai package reports the same values. The emitted attributes are unchanged. Assisted-by: Claude Opus 5
Krishnachaitanyakc
force-pushed
the
anthropic-cache-tokens-via-invocation-fields
branch
from
September 8, 2026 21:24
441736a to
be0d362
Compare
Merged
via the queue into
open-telemetry:main
with commit Sep 9, 2026
10f354d
74 of 75 checks passed
github-project-automation
Bot
moved this to Done
in GenAI Semantic Conventions and Instrumentation libraries
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Description
messages_extractorsdefined its owngen_ai.usage.cache_creation.input_tokensandgen_ai.usage.cache_read.input_tokensstring constants and wrote them straight intoinvocation.attributes, so the attribute names were pinned in this package rather than tracking the semantic conventions.InferenceInvocationalready carriescache_creation_input_tokensandcache_read_input_tokensfields for exactly this, and openai uses them (response_extractors.py:661).This sets those fields instead and drops the two local constants. The emitted attributes are unchanged.
Type of change
How has this been tested?
The existing cache token tests now assert against the semconv constants rather than the removed local ones.
anthropicsuite: 178 passed (latest), 171 passed with 7 skipped (oldest)anthropicconformance: 5 passedtox -e precommitandtox -e typecheckcleanChecklist