You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I got a bug report related to the datadog integration for httpx, which I've narrowed down to the faraday datadog integration (where faraday uses httpx as a backend, and httpx integration is itself turned off). In the report, the reporter shows a screenshot of a request span not propagating to the next service, which also integrates with datadog and has distributed tracing turned on.
While reproducing the issue, as per one of my last messages in the thread, I was able to spot some different behaviour in the faraday datadog integration (compared to the httpx integration) where, when under an active trace, the faraday span appears before the top span (see linked tests).
Something tells me that this is all related, and a bug perhaps in the faraday integration, as the span should have been appended instead, and this lack of ordering is what may cause the distributed tracing not to work in this case.
Any help appreciated.
Reproduction Code
No response
Configuration Block
No response
Error Logs
No response
Operating System
No response
How does Datadog help you?
No response
The text was updated successfully, but these errors were encountered:
1) Failure:
FaradayDatadogTest#test_faraday_datadog_distributed_headers_sampling_priority [integration_tests/datadog_helpers.rb:10]:
Expected false to be truthy.
2) Failure:
FaradayDatadogTest#test_faraday_datadog_distributed_headers_disabled [integration_tests/datadog_helpers.rb:10]:
Expected false to be truthy.
To bring all the information here, these two tests fail:
That's correct @marcotc ! FWIW you don't need to setup the docker based test env, you can just clone the repo, bundle install, and run each of those individually with bundle exec ruby -Itest integration_tests/faraday_datadog_test.rb --name /name_of_test/ .
is c.tracing.instrument :library being called for both :faraday and :httpx in the failing Faraday integration test?
No. In the integration_tests folder you'll find two test files: datadog_test.rb, which tests plain httpx with the datadog tracing adapter :httpx enabled; and faraday_datadog_test.rb, which tests faraday-using-httpx, with the datadog tracing adapter :faraday enabled.
Tracer Version(s)
2.9.0
Ruby Version(s)
ruby 3.3.5 (2024-09-03 revision ef084cc8f4) [x86_64-linux]
Relevent Library and Version(s)
faraday 2.12.2, httpx 1.4.0 (or gh-70 branch)
Bug Report
I got a bug report related to the datadog integration for httpx, which I've narrowed down to the faraday datadog integration (where faraday uses httpx as a backend, and httpx integration is itself turned off). In the report, the reporter shows a screenshot of a request span not propagating to the next service, which also integrates with datadog and has distributed tracing turned on.
While reproducing the issue, as per one of my last messages in the thread, I was able to spot some different behaviour in the faraday datadog integration (compared to the httpx integration) where, when under an active trace, the faraday span appears before the top span (see linked tests).
Something tells me that this is all related, and a bug perhaps in the faraday integration, as the span should have been appended instead, and this lack of ordering is what may cause the distributed tracing not to work in this case.
Any help appreciated.
Reproduction Code
No response
Configuration Block
No response
Error Logs
No response
Operating System
No response
How does Datadog help you?
No response
The text was updated successfully, but these errors were encountered: