Blog post for OpenTelemetry Generative AI updates #5575

drewby · 2024-11-09T00:40:17Z

Title: OpenTelemetry for Generative AI

This blog post introduces enhancements to OpenTelemetry specifically tailored for generative AI technologies, focusing on the development of Semantic Conventions and the Python Instrumentation Library.

Samples are in Python

SIG: GenAI Observability

Sponsors: @tedsuo @lmolkova

Closes: #5581

content/en/blog/2024/otel-generative-ai/index.md

Co-authored-by: Liudmila Molkova <[email protected]>

content/en/blog/2024/otel-generative-ai/index.md

lmolkova · 2024-11-09T01:10:37Z

content/en/blog/2024/otel-generative-ai/index.md

+from opentelemetry.instrumentation.openai_v2 import OpenAIInstrumentor
+
+OpenAIInstrumentor().instrument()
+


we probably need to add the code that configure otel SDK (or add a comment pointing to docs on how to set it up).

The code snippet:

# NOTE: OpenTelemetry Python Logs and Events APIs are in beta from opentelemetry import trace, _logs, _events from opentelemetry.sdk.trace import TracerProvider from opentelemetry.sdk._logs import LoggerProvider from opentelemetry.sdk._events import EventLoggerProvider from opentelemetry.sdk.trace.export import BatchSpanProcessor from opentelemetry.sdk._logs.export import BatchLogRecordProcessor from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter from opentelemetry.exporter.otlp.proto.grpc._log_exporter import OTLPLogExporter trace.set_tracer_provider(TracerProvider()) trace.get_tracer_provider().add_span_processor( BatchSpanProcessor(OTLPSpanExporter()) ) _logs.set_logger_provider(LoggerProvider()) _logs.get_logger_provider().add_log_record_processor(BatchLogRecordProcessor(OTLPLogExporter())) _events.set_event_logger_provider(EventLoggerProvider())

an existing example that does something similar - https://github.com/open-telemetry/opentelemetry-python/blob/stable/docs/examples/logs/example.py

/cc @lzchen @xrmx in case they have some suggestions

That's a lot of extra code. Should I link to that code example or is there a better doc?

let me add it to open-telemetry/opentelemetry-python-contrib#2988 and, once merged, let's just link it.

content/en/blog/2024/otel-generative-ai/index.md

content/en/blog/2024/otel-generative-ai/aspire_dashboard_trace.png

Co-authored-by: Liudmila Molkova <[email protected]>

theletterf

Cool one! Added some comments.

content/en/blog/2024/otel-generative-ai/index.md

codefromthecrypt · 2024-11-11T02:37:38Z

The below is about if we can make it zero code or not, and what's remaining to do so:

I want to run dotenv run -- opentelemetry-instrument python main.py and not have any telemetry code in the foreground or even noticed (zero code approach)

I have a similar example, I've tried locally, and I don't see a way to implicitly configure the logging provider, yet. I'm not sure if we want to make a hybrid to reduce the amount of code or just leave the explicit tracing and logging stuff in until logging can be env configured.

cc @anuraaga and @xrmx in case I got below wrong also

requirements

openai~=1.54.3

opentelemetry-sdk==1.28.1
opentelemetry-exporter-otlp-proto-http==1.28.1
opentelemetry-distro==0.49b1
opentelemetry-instrumentation-openai-v2==2.0b0

env

# Override default ENV variables for Ollama
OPENAI_BASE_URL=http://localhost:11434/v1
OPENAI_API_KEY=unused
CHAT_MODEL=qwen2.5:0.5b

# Set to false or remove to disable log events
OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true

OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318
OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
OTEL_SERVICE_NAME=python-opentelemetry-openai
OTEL_METRIC_EXPORT_INTERVAL=100
OTEL_BSP_SCHEDULE_DELAY=100

Best I could manage was to add hooks only for the log/event stuff

import os

from openai import OpenAI

# NOTE: OpenTelemetry Python Logs and Events APIs are in beta
from opentelemetry import _logs, _events
from opentelemetry.sdk._logs import LoggerProvider
from opentelemetry.sdk._events import EventLoggerProvider

from opentelemetry.sdk._logs.export import BatchLogRecordProcessor
from opentelemetry.exporter.otlp.proto.http._log_exporter import OTLPLogExporter

_logs.set_logger_provider(LoggerProvider())
_logs.get_logger_provider().add_log_record_processor(BatchLogRecordProcessor(OTLPLogExporter()))
_events.set_event_logger_provider(EventLoggerProvider())

def main():
    client = OpenAI()
    messages = [
        {
            "role": "user",
            "content": "Answer in up to 3 words: Which ocean contains the falkland islands?",
        },
    ]
    model = os.getenv("CHAT_MODEL", "gpt-4o-mini")
    chat_completion = client.chat.completions.create(model=model, messages=messages)
    print(chat_completion.choices[0].message.content)


if __name__ == "__main__":
    main()

then, I get a warning about overriding the event provider, but at least the events do show up

$ dotenv run -- opentelemetry-instrument python main.py
Overriding of current EventLoggerProvider is not allowed
Indian Ocean

codefromthecrypt · 2024-11-13T01:03:44Z

just to verify, is this issue blocking this PR?

not my call, but I think putting best foot forward is getting the "zero code" approach in the blog. Having otel infra code in the foreground of the project is not ideal, nor is having something look like it might work without code, then to try the demo and find out it doesn't work. I would go for best==zero code, and push for a patch. It is a small change to get us over the line and worth doing.,

codefromthecrypt · 2024-11-13T01:06:47Z

content/en/blog/2024/otel-generative-ai/index.md

+pip install opentelemetry-instrumentation-openai-v2
+```
+
+Then include the following code in your Python application:


this example is currently misleading, as it looks like it is zero code. The large setup logic is not visible in the blog. This is why I strongly don't agree with the optimization of time vs clean.

There are new facets people need to plumb for this to work as expected (e.g. log events). People in genai expect to see prompt/completion data. We should make it work (zero code approach) or make very visible the code required to have that happen.

svrnm · 2024-11-13T13:02:47Z

I see that the issue is fixed and pending a release of python SDK. @drewby I will leave it up to you if we wait for that or finalize this before.

.cspell/en-words.txt

content/en/blog/2024/otel-generative-ai/index.md

gyliu513 · 2024-11-13T18:46:24Z

content/en/blog/2024/otel-generative-ai/index.md

+participate in discussions. Explore the OpenTelemetry Python Contrib project,
+contribute code, or help shape observability for AI as it continues to evolve.
+More information can be found at the
+[Generative AI Observability project page](https://github.com/open-telemetry/community/blob/main/projects/gen-ai.md).


@drewby How about highlighting the current contributors at the end of the blog? This way, readers can quickly see who’s actively involved in this SIG.

Suggested change

[Generative AI Observability project page](https://github.com/open-telemetry/community/blob/main/projects/gen-ai.md).

[Generative AI Observability project page](https://github.com/open-telemetry/community/blob/main/projects/gen-ai.md), we now have contributors from [OpenLIT](https://openlit.io/), [Langtrace](https://www.langtrace.ai/), [Elastic](https://www.elastic.co/), [MicroSoft](https://www.microsoft.com/), [Traceloop](https://www.traceloop.com/), [IBM](https://www.ibm.com), [Scorecard](https://www.scorecard.io/), [Google](https://www.google.com/), [Amazon](https://aws.amazon.com/) etc., welcome to join the community!

Happy to include Scorecard (scorecard.io) as a supporter and adopter of GenAI OTEL!

@drewby How about highlighting the current contributors at the end of the blog? This way, readers can quickly see who’s actively involved in this SIG.

It's a good suggestion and I'd be delighted to include how we've had great collaboration and contribution across all these companies.

@svrnm I'm not sure how this fits with the community/blog guidelines?

codefromthecrypt · 2024-11-14T04:01:39Z

content/en/blog/2024/otel-generative-ai/index.md

+[Aspire Dashboard](https://learn.microsoft.com/dotnet/aspire/fundamentals/dashboard/standalone?tabs=bash)
+for local debugging.
+
+![Chat trace in Aspire Dashboard](aspire-dashboard-trace.png)


I noticed both of these traces are not from the sample code, as they have spans that it wouldn't produce. Is this ok?

drewby · 2024-11-14T04:13:25Z

update on #5575 (comment) based on feedback from @lzchen. We can pare down the code impact to a few lines, assuming you have env like below:

@codefromthecrypt I have updated the code example. Can you take a look?

codefromthecrypt

I think this needs more work and another person to double-check. I checked personally noticing the images don't match the main code. It failed in jaeger, but I don't think it is about jaeger as the failure also happens in otel-tui. I think the manual setup has a glitch in it and we shouldn't publish this blog until the code here matches the images and also it doesn't produce exceptions exporting logs. If it is my mistake, please let me know what I did wrong.

I'm wondering if someone else can double-check this example as-is because it failed for me against jaeger v2 like so:

$ OPENAI_API_KEY=not-tellin python main.py
Overriding of current EventLoggerProvider is not allowed
Exception while exporting logs.
Traceback (most recent call last):
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/sdk/_logs/_internal/export/__init__.py", line 307, in _export_batch
    self._exporter.export(self._log_records[:idx])  # type: ignore
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/_log_exporter/__init__.py", line 111, in export
    return self._export(batch)
           ^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/exporter.py", line 300, in _export
    request=self._translate_data(data),
            ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/_log_exporter/__init__.py", line 108, in _translate_data
    return encode_logs(data)
           ^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 37, in encode_logs
    return ExportLogsServiceRequest(resource_logs=_encode_resource_logs(batch))
                                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 71, in _encode_resource_logs
    pb2_log = _encode_log(sdk_log)
              ^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 57, in _encode_log
    body=_encode_value(log_data.log_record.body),
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/__init__.py", line 90, in _encode_value
    raise Exception(f"Invalid type {type(value)} of value {value}")
Exception: Invalid type <class 'NoneType'> of value None

I used the following requirements.txt

openai~=1.54.3

# 1.28.1 is required for Log Events API/SDK
opentelemetry-sdk~=1.28.1
opentelemetry-exporter-otlp-proto-grpc~=1.28.1
opentelemetry-distro~=0.49b1
opentelemetry-instrumentation-openai-v2~=2.0b0

and I used exactly the code as listed here:

from opentelemetry import trace, _logs, _events
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk._logs import LoggerProvider
from opentelemetry.sdk._events import EventLoggerProvider

from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.sdk._logs.export import BatchLogRecordProcessor
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
from opentelemetry.exporter.otlp.proto.grpc._log_exporter import OTLPLogExporter

trace.set_tracer_provider(TracerProvider())
trace.get_tracer_provider().add_span_processor(
    BatchSpanProcessor(OTLPSpanExporter())
)

_logs.set_logger_provider(LoggerProvider())
_logs.get_logger_provider().add_log_record_processor(BatchLogRecordProcessor(OTLPLogExporter()))
_events.set_event_logger_provider(EventLoggerProvider())

from openai import OpenAI
from opentelemetry.instrumentation.openai_v2 import OpenAIInstrumentor

OpenAIInstrumentor().instrument()

client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Write a short poem on OpenTelemetry."}],
)

For my jaeger setup I needed to use the following config because you can no longer listen on all points via ENV variable. I actually just wanted to override the listen ports, but don't know how to do that so I merged the receiver config with a file in the repo. cc @yurishkuro

service:
  extensions: [jaeger_storage, jaeger_query, remote_sampling, healthcheckv2, expvar]
  pipelines:
    traces:
      receivers: [otlp, jaeger, zipkin]
      processors: [batch]
      exporters: [jaeger_storage_exporter]
  telemetry:
    resource:
      service.name: jaeger
    metrics:
      level: detailed
      address: 0.0.0.0:8888
    logs:
      level: info
    # TODO Initialize telemetery tracer once OTEL released new feature.
    # https://github.com/open-telemetry/opentelemetry-collector/issues/10663

extensions:
  jaeger_query:
    storage:
      traces: some_storage

  jaeger_storage:
    backends:
      some_storage:
        memory:
          max_traces: 100000

  remote_sampling:
    # We can either use file or adaptive sampling strategy in remote_sampling
    file:
      path: ./cmd/jaeger/sampling-strategies.json
    # adaptive:
    #   sampling_store: some_store
    #   initial_sampling_probability: 0.1
    http:
    grpc:

  healthcheckv2:
    use_v2: true
    http:
      endpoint: "0.0.0.0:13133"
    grpc:

  expvar:
    port: 27777

receivers:
  otlp:
    protocols:
      grpc:
        endpoint: 0.0.0.0:4317
      http:
        endpoint: 0.0.0.0:4318

  jaeger:
    protocols:
      grpc:
      thrift_binary:
      thrift_compact:
      thrift_http:

  zipkin:

processors:
  batch:

exporters:
  jaeger_storage_exporter:
    trace_storage: some_storage

content/en/blog/2024/otel-generative-ai/index.md

yurishkuro · 2024-11-14T04:51:04Z

content/en/blog/2024/otel-generative-ai/index.md

+![Chat trace in Aspire Dashboard](aspire-dashboard-trace.png)
+
+Here is a similar trace captured in
+[Jaeger](https://www.jaegertracing.io/docs/next-release-v2/getting-started/#running):


Suggested change

[Jaeger](https://www.jaegertracing.io/docs/next-release-v2/getting-started/#running):

[Jaeger](https://www.jaegertracing.io/docs/1.63/getting-started/#all-in-one):

drewby · 2024-11-14T05:05:47Z

I think this needs more work and another person to double-check. I checked personally noticing the images don't match the main code. It failed in jaeger, but I don't think it is about jaeger as the failure also happens in otel-tui. I think the manual setup has a glitch in it and we shouldn't publish this blog until the code here matches the images and also it doesn't produce exceptions exporting logs. If it is my mistake, please let me know what I did wrong.

I'm wondering if someone else can double-check this example as-is because it failed for me against jaeger v2 like so:

I've updated the code sample to something similar to yours. Would love to reduce the amount of code, but also want to make sure it works.

Do you think we need to include the requirements and/or pip install instructions for openai, opentelemetry, etc?

codefromthecrypt · 2024-11-14T05:14:07Z

So, the last code I pasted works fine, but only if you also set OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true

Through trial and error, what I figured out is that if you have the boilerplate python to configure logging, but you don't also set OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true, then the stack trace below comes out.
This seems like a bug as it should fallback safely right?

Overriding of current EventLoggerProvider is not allowed
Exception while exporting logs.
Traceback (most recent call last):
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/sdk/_logs/_internal/export/__init__.py", line 307, in _export_batch
    self._exporter.export(self._log_records[:idx])  # type: ignore
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/_log_exporter/__init__.py", line 111, in export
    return self._export(batch)
           ^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/exporter.py", line 300, in _export
    request=self._translate_data(data),
            ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/grpc/_log_exporter/__init__.py", line 108, in _translate_data
    return encode_logs(data)
           ^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 37, in encode_logs
    return ExportLogsServiceRequest(resource_logs=_encode_resource_logs(batch))
                                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 71, in _encode_resource_logs
    pb2_log = _encode_log(sdk_log)
              ^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/_log_encoder/__init__.py", line 57, in _encode_log
    body=_encode_value(log_data.log_record.body),
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/adriancole/oss/otel-ollama/examples/python/opentelemetry/openai/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/__init__.py", line 90, in _encode_value
    raise Exception(f"Invalid type {type(value)} of value {value}")
Exception: Invalid type <class 'NoneType'> of value None

codefromthecrypt · 2024-11-14T05:22:19Z

so current status is I can get the example working on otel-tui, but only if I export that variable:

$ OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true  OPENAI_API_KEY=sk-not_telling python main.py
Overriding of current EventLoggerProvider is not allowed

That said, I get a failure on jaeger v2, but @yurishkuro might want to double-check me as maybe I messed something up.

$ OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true  OPENAI_API_KEY=sk-not_telling python main.py
Overriding of current EventLoggerProvider is not allowed
Failed to export logs to localhost:4317, error code: StatusCode.UNIMPLEMENTED

codefromthecrypt

@drewby FYI I re-pasted the code from your latest commit and it doesn't change the error when content capture is set to false.

The main thing that's a discrepancy are the traces. The code only produces a single span, and both examples have multiple spans. Probably least confusing to reproduce the traces based on the code in the example, so people don't wonder why their trace has only one vs the three in the images. WDYT?

on pip install vs requirements, it is fine either way as long as the end results work. The main issue is that pip install is only doing the opentelemetry-instrumentation-openai-v2 part, so it is missing the other dependencies I mentioned in my requirements.txt

Make sense?

As mentioned earlier, best case is we can use the zero code bootstrap thing which reduces the amount of commands and visible code. So the above are mostly around the manual trace infra code approach.

codefromthecrypt · 2024-11-14T05:23:03Z

content/en/blog/2024/otel-generative-ai/index.md

@@ -122,29 +122,47 @@ appropriate:
 OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318
 OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
 OTEL_SERVICE_NAME=python-opentelemetry-openai
+OPENAI_API_KEY=<replace_with_your_openai_api_key>

 # Set to false or remove to disable log events
 OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true


main thing here is that if you set this to false, it will crash (at least for me)

@lzchen @xrmx can one of you double-check the stack trace which happens when this is set to false or unset? I'm not sure if this is already fixed in main or not.

#5575 (comment)

open-telemetry/opentelemetry-python#4276 or similar will fix this, though not sure if all backends expect a log with no body or not.

codefromthecrypt · 2024-11-14T05:25:32Z

content/en/blog/2024/otel-generative-ai/index.md

+from opentelemetry._events import set_event_logger_provider
+from opentelemetry.sdk.trace.export import SimpleSpanProcessor
+from opentelemetry.sdk._logs.export import SimpleLogRecordProcessor
+from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter


while grpc is ok I'm not sure if all vendors support that. Also it requires platform deps to build (gcc which is slow to install). Should we switch this to http instead?

I see now, it is because aspire only supports grpc https://learn.microsoft.com/en-us/dotnet/aspire/fundamentals/dashboard/standalone?tabs=bash#configure-opentelemetry-sdk

codefromthecrypt · 2024-11-14T05:37:10Z

content/en/blog/2024/otel-generative-ai/index.md

+appropriate:
+
+```bash
+OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318


TL;DR; IMHO whatever we paste in a bash codeblock, people will expect to work by pasting on their own. This doesn't work, yet even if we fix the pip deps due to not exporting the variables.

maybe worth removing these two OTEL_EXPORTER variables and leave it to defaults based on the dependency like opentelemetry-exporter-otlp-proto-grpc or opentelemetry-exporter-otlp-proto-http. wdyt?

Then, we have only the service name, api key and capture content ones. Also, these need to be exported or set inline. If you don't the python command won't work later. Many people put these into a .env file and use dotenv to load it.

e.g.

$ pip install "python-dotenv[cli]" $ dotenv run -- python main.py

However, we can also just inline them like

$ OTEL_SERVICE_NAME=python-opentelemetry-openai OPENAI_API_KEY=sk-not_tellin OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true python main.py

codefromthecrypt · 2024-11-14T08:07:30Z

follow-up for jaeger jaegertracing/documentation#778

codefromthecrypt · 2024-11-15T01:36:54Z

here's example code which would stay in draft until the upcoming release of 1.28.2 and any feedback open-telemetry/opentelemetry-python-contrib#3006

codefromthecrypt

Here's some food for thought on jaeger which only accepts span data. Since the demo can send logs and traces, configuration care is needed.

Notably, OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=false doesn't disable log export, rather it just pares down the log events such that they have no body. So, this means you will still get errors as it attempts to send log messages to jaeger where they aren't implemented.

One way is to make an ENV that comments out everything here, so that it works with trace-only backends, but still shows how to use it in aspire. That said, this makes more steps for folks to see the prompt/completion data and that's a big part of this instrumentation. Another way is to leave the demo setup in a way that is optimized for collectors that accept logs, and mention this implies commenting this part out if yours doesn't.

# Comment out the following if your backend doesn't support OTLP logs
OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=true
OTEL_PYTHON_LOGGING_AUTO_INSTRUMENTATION_ENABLED=true
OTEL_LOGS_EXPORTER=otlp_proto_http

Regardless, when the code settles someone should give second eyes to the Aspire approach which appears to accept all signals. I've not yet.

drewby added 4 commits October 25, 2024 15:46

Initial draft

5c062d9

Merge branch 'open-telemetry:main' into drewby/genai_blog

e4735c4

Add screenshots

5c9a04b

Update link for Python repo

c3d7c9e

drewby requested a review from a team as a code owner November 9, 2024 00:40

opentelemetrybot requested a review from a team November 9, 2024 00:40

github-actions bot added the blog label Nov 9, 2024

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

fix spelling error

db18921

Co-authored-by: Liudmila Molkova <[email protected]>

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

lmolkova reviewed Nov 9, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/aspire_dashboard_trace.png Outdated Show resolved Hide resolved

drewby and others added 9 commits November 9, 2024 10:52

Add link to Aspire Dashboard

67f2bc9

Co-authored-by: Liudmila Molkova <[email protected]>

Add link to Jaeger

6a51e0c

Co-authored-by: Liudmila Molkova <[email protected]>

Add words to cspell

c66f6b4

Updates from PR review

fdef792

Rename files

35c2487

Pin version

1796b6e

Update linting/spelling error

8dd9a11

Fix format error

4159b7d

Updates to library intro and metric section

480235b

theletterf reviewed Nov 10, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

drewby added 3 commits November 11, 2024 08:01

Update introduction

aef7577

Use headers instead of bold

16877ad

Fix formatting

c7e3ca9

opentelemetrybot requested a review from a team November 10, 2024 23:01

codefromthecrypt reviewed Nov 13, 2024

View reviewed changes

svrnm reviewed Nov 13, 2024

View reviewed changes

.cspell/en-words.txt Outdated Show resolved Hide resolved

svrnm reviewed Nov 13, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

dasiths reviewed Nov 13, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Show resolved Hide resolved

gyliu513 reviewed Nov 13, 2024

View reviewed changes

drewby added 2 commits November 14, 2024 12:58

Merge branch 'open-telemetry:main' into drewby/genai_blog

be9608f

Move words from cspell to article header

638e3b4

opentelemetrybot requested a review from a team November 14, 2024 03:58

codefromthecrypt reviewed Nov 14, 2024

View reviewed changes

drewby added 3 commits November 14, 2024 04:02

Fix lint/link errors

0303e36

Update code sample

a1489fb

fix format

730e1c1

codefromthecrypt reviewed Nov 14, 2024

View reviewed changes

content/en/blog/2024/otel-generative-ai/index.md Outdated Show resolved Hide resolved

yurishkuro reviewed Nov 14, 2024

View reviewed changes

Update code sample

64cd559

codefromthecrypt reviewed Nov 14, 2024

View reviewed changes

codefromthecrypt mentioned this pull request Nov 15, 2024

Add OpenAI example open-telemetry/opentelemetry-python-contrib#3006

Draft

11 tasks

codefromthecrypt reviewed Nov 15, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Blog post for OpenTelemetry Generative AI updates #5575

Blog post for OpenTelemetry Generative AI updates #5575

drewby commented Nov 9, 2024 •

edited

Loading

lmolkova Nov 9, 2024

drewby Nov 9, 2024

lmolkova Nov 10, 2024

theletterf left a comment

codefromthecrypt commented Nov 11, 2024

codefromthecrypt commented Nov 13, 2024

codefromthecrypt Nov 13, 2024 •

edited

Loading

svrnm commented Nov 13, 2024

gyliu513 Nov 13, 2024 •

edited

Loading

Rutledge Nov 13, 2024 •

edited

Loading

drewby Nov 14, 2024

codefromthecrypt Nov 14, 2024

drewby commented Nov 14, 2024

codefromthecrypt left a comment

yurishkuro Nov 14, 2024 •

edited

Loading

drewby commented Nov 14, 2024

codefromthecrypt commented Nov 14, 2024

codefromthecrypt commented Nov 14, 2024

codefromthecrypt left a comment

codefromthecrypt Nov 14, 2024

codefromthecrypt Nov 14, 2024

codefromthecrypt Nov 14, 2024

codefromthecrypt Nov 14, 2024 •

edited

Loading

codefromthecrypt Nov 15, 2024

codefromthecrypt Nov 14, 2024

codefromthecrypt commented Nov 14, 2024

codefromthecrypt commented Nov 15, 2024

codefromthecrypt left a comment

		from opentelemetry.instrumentation.openai_v2 import OpenAIInstrumentor

		OpenAIInstrumentor().instrument()

	[Jaeger](https://www.jaegertracing.io/docs/next-release-v2/getting-started/#running):
	[Jaeger](https://www.jaegertracing.io/docs/1.63/getting-started/#all-in-one):

Blog post for OpenTelemetry Generative AI updates #5575

Are you sure you want to change the base?

Blog post for OpenTelemetry Generative AI updates #5575

Conversation

drewby commented Nov 9, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

theletterf left a comment

Choose a reason for hiding this comment

codefromthecrypt commented Nov 11, 2024

codefromthecrypt commented Nov 13, 2024

codefromthecrypt Nov 13, 2024 • edited Loading

Choose a reason for hiding this comment

svrnm commented Nov 13, 2024

gyliu513 Nov 13, 2024 • edited Loading

Choose a reason for hiding this comment

Rutledge Nov 13, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

drewby commented Nov 14, 2024

codefromthecrypt left a comment

Choose a reason for hiding this comment

yurishkuro Nov 14, 2024 • edited Loading

Choose a reason for hiding this comment

drewby commented Nov 14, 2024

codefromthecrypt commented Nov 14, 2024

codefromthecrypt commented Nov 14, 2024

codefromthecrypt left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codefromthecrypt Nov 14, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codefromthecrypt commented Nov 14, 2024

codefromthecrypt commented Nov 15, 2024

codefromthecrypt left a comment

Choose a reason for hiding this comment

drewby commented Nov 9, 2024 •

edited

Loading

codefromthecrypt Nov 13, 2024 •

edited

Loading

gyliu513 Nov 13, 2024 •

edited

Loading

Rutledge Nov 13, 2024 •

edited

Loading

yurishkuro Nov 14, 2024 •

edited

Loading

codefromthecrypt Nov 14, 2024 •

edited

Loading