[ 
https://issues.apache.org/jira/browse/CAMEL-23369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18147769#comment-18147769
 ] 

Federico Mariani commented on CAMEL-23369:
------------------------------------------

h3. Status, 19 Nov 2026

*Done*
* 4.23.0, mostly under CAMEL-23861 and its follow-ups: 
{{camel-ai-observability}} (opt-in) records GenAI spans with the {{gen_ai.*}} 
attributes and the Micrometer metrics {{gen_ai.client.operation}} (timer) and 
{{gen_ai.client.token.usage}} (counter). It covers {{openai}} (chat-completion 
with streaming, responses, embeddings, moderation; CAMEL-24559), 
{{langchain4j-chat}}, {{langchain4j-agent}}, {{langchain4j-embeddings}} and 
{{spring-ai-chat}}. Related: Micrometer Observation (CAMEL-24561), error 
metadata (CAMEL-24551), Camel Main and Spring Boot options (CAMEL-24386, 
CAMEL-24555), Camel CLI dependencies (CAMEL-24566).
* 4.24.0: the {{basic}} / {{detailed}} / {{full}} levels with agent runs, tool 
calls and messages, and no double recording of the calls already traced by 
Spring AI, quarkus-langchain4j or the OpenTelemetry Java agent (CAMEL-27131). 
The implementation is found without reflection (CAMEL-27155). Spring Boot: 
CAMEL-27132, CAMEL-27133, CAMEL-27134.
* Cached input tokens ({{gen_ai.usage.cache_read.input_tokens}}) for {{openai}} 
chat-completion, at level {{detailed}}.

*Open*
* CAMEL-27142: {{openai}} responses at {{detailed}} and {{full}}. PR 
https://github.com/apache/camel/pull/30326 needs a rebase, because #30339 
removed {{GenAiObservation.makeCurrent()}}.
* CAMEL-27143: read the model name, provider and error status without 
reflection (native images).
* CAMEL-27135: the camel-jbang views for {{detailed}} and {{full}}.
* Camel Quarkus extension: https://github.com/apache/camel-quarkus/issues/9518 
(no PR yet).
* Related: CAMEL-27139 (span names), CAMEL-27140 (session and user ids on every 
span).

*Not covered yet, no ticket*
* {{camel-aws-bedrock}}, {{camel-google-vertexai}} and 
{{camel-ibm-watsonx-ai}}, listed in this issue, set the token counts as headers 
({{CamelAwsBedrockConverseUsage}}, {{CamelGoogleVertexAIPromptTokenCount}} / 
{{CandidatesTokenCount}}, {{CamelIBMWatsonxAiInputTokenCount}} / 
{{OutputTokenCount}}) but record no GenAI span or metric. CAMEL-24559 proposed 
them and was resolved with the {{openai}} operations only.
* Cache write and reasoning tokens are not recorded by any component. Cached 
input tokens are recorded only by {{openai}} chat-completion, on the span, not 
as a metric.
* The GenAI metrics go through Micrometer only: an application with 
{{camel-opentelemetry-metrics}} and no Micrometer gets the GenAI spans but no 
GenAI metrics.
* The GenAI conventions moved to 
https://github.com/open-telemetry/semantic-conventions-genai (semconv 1.45.0). 
That repository has no release yet, and its main branch renames metrics and 
attributes, for example {{gen_ai.client.token.usage}} becomes 
{{gen_ai.client.inference.usage.*}} counters, and {{cache_creation}} becomes 
{{cache_write}}. Camel records {{gen_ai.system}}, deprecated in favour of 
{{gen_ai.provider.name}}, at every level, and {{gen_ai.provider.name}} only 
from {{detailed}}.

_Claude Code on behalf of Croway_


> Add OpenTelemetry GenAI observability (traces and metrics) for AI components
> ----------------------------------------------------------------------------
>
>                 Key: CAMEL-23369
>                 URL: https://issues.apache.org/jira/browse/CAMEL-23369
>             Project: Camel
>          Issue Type: Improvement
>    Affects Versions: 4.20.0
>            Reporter: Federico Mariani
>            Priority: Major
>
> Currently none of the Camel AI components (camel-openai, 
> camel-langchain4j-chat, camel-langchain4j-embeddings, camel-spring-ai-chat, 
> camel-google-vertexai, camel-aws-bedrock, camel-ibm-watsonx-ai) are 
> instrumented by the _camel-telemetry_ tracing framework or the metrics 
> frameworks (camel-opentelemetry-metrics, camel-micrometer).
> Similarly, the existing EventNotifier-based metrics (exchange duration, 
> in-flight gauges) don't capture AI-specific data like token counts.           
>                                                                               
>        
>                   
> The https://opentelemetry.io/docs/specs/semconv/gen-ai/ define standard span 
> attributes (gen_ai.operation.name, gen_ai.provider.name, 
> gen_ai.request.model, gen_ai.usage.input_tokens, gen_ai.usage.output_tokens, 
> gen_ai.response.finish_reasons, etc.) and metrics (gen_ai.client.token.usage 
> histogram, gen_ai.client.operation.duration histogram) for LLM operations. 
> Camel AIs components already expose token counts and response metadata as 
> Camel message headers, but this data does not flow into distributed traces or 
> metrics.
> We might improve the current camel AI implementations, and expose all the 
> important metrics if not already exposed (for example, cache read and cache 
> write tokens)



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

Reply via email to