Thank you so much for looking into this.  Sorry I had traveled for *Xmas*.

On Sat, Dec 25, 2021 at 8:39 PM Clebert Suconic <clebert.suco...@gmail.com>
wrote:

> At this point I’m just helping Brenda to make progress. After we do some
> progress we can look into making agnostic.
>
> On Fri, Dec 24, 2021 at 9:27 AM michael.andre.pearce
> <michael.andre.pea...@me.com.invalid> wrote:
>
> > Tbh it should be vendor agnostic neither zipkin or jaeger this is the
> > point of open telemetry. We shouldn't be packaging up zipkin or jaeger
> > specific.Im sure every org has their own tracing vendor solution, as such
> > just should be done so that we supply the plugin to api but users should
> > add the appropriate agent as they need. Sent from my Galaxy
> > -------- Original message --------From: Clebert Suconic <
> > clebert.suco...@gmail.com> Date: 23/12/2021  14:08  (GMT+00:00) To:
> > dev@activemq.apache.org Subject: Re: [DISCUSS] Open Telemetry with
> > ActiveMQ Jaeger will accept Zipkin packets.. if you write zipkin it
> should
> > workwith either Jaeger or Zipkin... so I think it's very nice for
> > aprototype.I think we could start working on making it ready for PR...
> > create theappropriate sub-modules and proper tests.Is there any API that
> > could receive tracing... I wonder how we wouldwrite testcases...As for
> > making it ready we could create the sub-module asartemis-open-tracing....
> > and get the URI for open tracing from theconfiguration.xmlOn Tue, Dec 21,
> > 2021 at 2:58 PM Nabwegamo Brenda <brenser...@gmail.com> wrote:>> Hello
> > Everyone!>> I am an Outreachy Intern implementing distributed tracing
> with
> > opentelemetry> <http://opentelemetry.io/> in activemq-artemis under
> > active mentorship> of @Clebert> Suconic>> *How we wish to Implement
> this.*>
> > The current goal for now is writing a Broker-plugin that will send data
> > for> every message sent to opentelemetry servers either using the API
> > and/or SDK> with exporters like Jeager <https://www.jaegertracing.io/>,
> > Zipkin> <https://zipkin.io/>, OTLP etc, and packaging it nicely>> I have
> > opened a PR of a working prototype of how this might work.> <
> >
> https://github.com/nbrendah/activemq-artemis/commit/a57adecdd56aaf7c5d46bfd065a40e010d29ed0f
> >>>
> > In little explanation of the code,>>    - On calling the  initZipkin()
> we
> > are initializing and managing the>    Opentelemetry SDK but this should
> > take place only once, like when the>    connection is started, or
> > otherwise, we might get some errors when the SDK>    configuration is
> > called again though some data will be received on Zipkin.>    - Each
> time a
> > message is created, before sending it, we create a span>    and set its
> > attributes (if any) and finally end the span.>> *What needs improvement
> > (not well implemented)*>>    - I need to come up with a way of executing
> > initZipkin()  once like at>    connection creation or something like that
> > and storing the tracer so that>    we can access it through
> > BrokerPlugin.beforeSend(...)  method.>    - Each time we need to create a
> > span, we can just get it from the>    session or message and then create
> > it.>>> *What I have failed to implement for now.*>>    - Passing a tracer
> > created on the static main method to the overridden>
> > BrokerPlugin.beforeSend(...)  so that we can create a span from the>
> > tracer before sending each message>    - At first I tried exporting the
> > spans to Jeager>    <https://www.jaegertracing.io/> using the
> > Jeager-example>    <
> >
> https://github.com/open-telemetry/opentelemetry-java/tree/main/examples/jaeger
> >>
> > but i wasn't receiving data for the last two weeks but Zipkin worked
> > fine>    on the first attempt.  However, i think Jeager and Zipkin work
> > almost the>    same according to me>>> *Questions  *>>    1. Which tool
> > would you personally go with. Jeager or Zipkin?>    2. Where can we
> > implement OpenTelemetry SDK initiation so that the>    initiation runs
> only
> > once?>    3. How best can we store a tracer such that it can be accessed
> > before a>    message is sent.-- Clebert Suconic
>
> --
> Clebert Suconic
>

Reply via email to