Jaeger: Ruby / Rails client

[jalberto]

Is it there any work to add a ruby client?

If it is not, Is it possible to use zipkin-ruby with jaeger?

[yurishkuro]

There is currently no Jaeger client in Ruby. A contribution would be awesome (but I don't think Ruby opentracing API is finalized).

I don't know how zipkin-ruby sends the traces out of process, but if it's in zipkin thrift format over HTTP, then Jaeger backend can accept them.

[jalberto]

do you mean this one? https://github.com/opentracing/opentracing-ruby

[yurishkuro]

Yes. I didn't follow it too closely as I don't know Ruby that well, but it looks like it might be in the "released" state already. @tedsuo do you know what the state of Ruby API?

[jalberto]

I would be able to help in ruby side of things :)

On Thu, 13 Jul 2017 at 16:46 Yuri Shkuro [e-mail] wrote:

Yes. I didn't follow it too closely as I don't know Ruby that well, but it
looks like it might be in the "released" state already. @tedsuo
https://github.com/tedsuo do you know what the state of Ruby API?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/uber/jaeger/issues/268#issuecomment-315099851, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAGGV_Mg0Gdf_MsruW-V3CDt1MLkkVLEks5sNi21gaJpZM4OXD-s
.

[yurishkuro]

So as a guideline, the following is a rough outline for getting a new client out:

1. setup new project, travis build, code coverage, release publishing, etc.
2. implement basic tracer & span functionality with a simple 100% sampling & a configuration mechanism
3. implement remote reporter with abstract sender
4. implement UDP or HTTP sender or both
5. implement other samplers (probabilistic, rate limiting)
6. implement remote sampler that pulls sampling strategy from the agent
7. implement adaptive sampler
8. optionally other features like baggage whitelisting

Steps 5-8 could be added later, 1-4 are needed to get a minimal viable client.

[jalberto]

I will need to get familiar with opentrace first

On Thu, 13 Jul 2017 at 17:13 Yuri Shkuro [e-mail] wrote:

So as a guideline, the following is a rough outline for getting a new
client out:

1. setup new project, travis build, code coverage, release publishing,
   etc.
2. implement basic tracer & span functionality with a simple 100%
   sampling & a configuration mechanism
3. implement remote reporter with abstract sender
4. implement UDP or HTTP sender or both
5. implement other samplers (probabilistic, rate limiting)
6. implement remote sampler that pulls sampling strategy from the agent
7. implement adaptive sampler
8. optionally other features like baggage whitelisting

Steps 5-8 could be added later, 1-4 are needed to get a minimal viable
client.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/uber/jaeger/issues/268#issuecomment-315108578, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAGGV0vGK-CIFBCycQs0C16j2QCz5Plwks5sNjQrgaJpZM4OXD-s
.

[tedsuo]

@yurishkuro I've been concentrating on Java/Python/C++11, was going to look at ruby after python. The ruby API will need context management at some point just like the others, but otherwise I can say it is in production and working at LightStep.

[indrekj]

I published a very basic ruby client some time ago: https://github.com/salemove/jaeger-client-ruby

[jpkrohling]

@indrekj would you be interested in having the Ruby client as part of this organization?

[danielmbarlow]

For info I use opencensus-ruby in production to communicate with Jaeger via opencensus-ruby-exporter-jaeger and I'm pretty happy with it.

Notes:

• OpenTracing and OpenCensus are merging into OpenTelemetry
• If you want Sidekiq integration this is what we're using