Reviewing files that changed from the base of the PR and between 5b13734 and f255fda.

• internal/dev/dev.go (1 hunks)

• GitHub Check: Test CLI Upgrade Path (windows-latest)

internal/dev/dev.go (1)

126-130:

Details

❓ Verification inconclusive

Good implementation following existing patterns, but verify the environment variable specifics.

The conditional logic follows the established pattern for language-specific environment setup. The purpose is clear from the comment and aligns with the PR objective to disable conflicting OpenTelemetry instrumentation.

However, please verify that:

1. The environment variable names are correct for the intended libraries
2. The hardcoded list "langchain,openai" covers all necessary instrumentations
3. This change doesn't break existing Python projects that might depend on these instrumentations

---

🏁 Script executed:

#!/bin/bash
# Verify the correct environment variable names for disabling Python OTEL instrumentation
rg -A 3 -B 3 "TRACELOOP_TRACE_CONTENT|OTEL_PYTHON_DISABLED_INSTRUMENTATIONS" --type py

Length of output: 237

---

🌐 Web query:

What are the correct environment variable names for disabling OpenTelemetry Python instrumentation, specifically for traceloop and the OTEL_PYTHON_DISABLED_INSTRUMENTATIONS setting?

💡 Result:

To disable specific instrumentations in OpenTelemetry Python, you can use the OTEL_PYTHON_DISABLED_INSTRUMENTATIONS environment variable. This variable accepts a comma-separated list of package names corresponding to the instrumentations you wish to exclude. For example, to disable instrumentations for redis, kafka-python, and grpc, you would set:

export OTEL_PYTHON_DISABLED_INSTRUMENTATIONS=redis,kafka-python,grpc

This setting ensures that the OpenTelemetry Python agent does not instrument these specific packages, preventing the collection of telemetry data from them. (opentelemetry.io)

Regarding the TRACELOOP_TELEMETRY environment variable, it is used to control telemetry data collection for the Traceloop SDK, which is separate from OpenTelemetry. To opt out of telemetry collection in Traceloop, you can set:

export TRACELOOP_TELEMETRY=FALSE

This setting disables the collection of anonymous usage information by Traceloop. (traceloop.com)

It's important to note that TRACELOOP_TELEMETRY does not affect OpenTelemetry's telemetry data collection. To disable telemetry in OpenTelemetry Python, you would typically set the OTEL_SDK_DISABLED environment variable to true:

export OTEL_SDK_DISABLED=true

This setting disables the OpenTelemetry SDK for all signals, effectively turning off telemetry data collection. (opentelemetry.io)

In summary, to disable specific instrumentations in OpenTelemetry Python, use the OTEL_PYTHON_DISABLED_INSTRUMENTATIONS environment variable. To disable telemetry data collection in Traceloop, use the TRACELOOP_TELEMETRY environment variable. To disable all telemetry in OpenTelemetry Python, use the OTEL_SDK_DISABLED environment variable.

Citations:

• 1: https://opentelemetry.io/docs/zero-code/python/configuration/?utm_source=openai
• 2: https://www.traceloop.com/docs/openllmetry/privacy/telemetry?utm_source=openai
• 3: https://opentelemetry.io/docs/specs/otel/configuration/sdk-environment-variables/?utm_source=openai

---

Verify Python OTEL environment variables and instrumentation names

The Python-specific branch aligns with the existing pattern for disabling automatic instrumentation, but please double-check the following before merging:

• The Traceloop environment variable. Upstream docs reference TRACELOOP_TELEMETRY=FALSE to opt out of Traceloop SDK telemetry, not TRACELOOP_TRACE_CONTENT.
• The OpenTelemetry instrumentation names. Ensure langchain,openai exactly match the package names used by the OpenTelemetry Python agent.
• Run manual tests against existing Python projects to confirm these settings don’t inadvertently disable needed instrumentation.