Closed mjgorman closed 5 years ago
Hey @mjgorman
If I understand correctly, you want the INCR
and GET
to automatically become children of your redis-span
Span, correct? If that's the case, this is part of an ongoing effort, as part of adopting the OT 2.0 API (which adds active Span
propagation & handling): https://github.com/opentracing-contrib/python-redis/tree/v1.0.0
The ETA for this effort to be finished is 1-2 weeks ;)
@carlosalberto great to hear. Glad it wasn't just something I was doing incorrectly.
Hey @mjgorman
I just released the 1.0.0
version, which should fix this. Simply remember to update your code to look like this:
with tracer.start_active_span('redis-span') as scope:
# Here your Redis commands and other related code.
# As 'redis_span' is the active Span, all new Spans
# under this block will become implicit children of it.
Let me know ;)
Switching start_span to start_active_span throw an error.
File "02b.py", line 36, in do_redis_things
with tracer.start_active_span('redis-span') as scope:
File "/Users/michael.gorman/sites/opentracing/python/venv/lib/python3.7/site-packages/opentracing/tracer.py", line 126, in start_active_span
return self._noop_scope
AttributeError: 'Tracer' object has no attribute '_noop_scope'
Hey @mjgorman
Thanks for the answer. So you are using Jaeger, right? It seems they still haven't released a version that implements the OT 2.0 api (which automatically detects the parent/child relationship).
As far as I know, LightStep, DataDog and Instana clients support this already ;) (I can poke the Jaeger guys and maybe help them to have this released, too). Let me know.
@carlosalberto Ah I see. Thank you for the informative response. I'll take it up with them.
I'm doing a little sample POC of OpenTracing for our team and I am getting the traces for redis, but they're separate from my spans which are running those redis commands. I'm setting a tag and that all works, but in jaeger they just come out as separate traces making it sort of useless.