also hi, sorry ii missed the mtg, i am a bit behind at the 9-5 but i willl try to get eyes on the dd related instrumentation ports and the rails work this week
Hello,
I've just updated my "demo" application to the release 0.8.0 and I've faced some issues. I was using
trace_id = OpenTelemetry::Instrumentation::Sinatra::Instrumentation.instance.tracer.current_span.context.trace_id.unpack('H*')[0]to get a trace_id but structure has been changed and right now current_span is not reachable. Could you please tell me how can I get current span or point some docs? And one more thing, how can I add an event to the span?
4 replies
@mwear et al - while transitioning apps to B3, would it make sense to expand the block below to also use w3c?
require 'opentelemetry-sdk'
require 'opentelemetry-propagator-b3'
OpenTelemetry::SDK.configure do |c|
c.http_extractors = [OpenTelemetry::Propagator::B3::Single.rack_extractor, OpenTelemetry::Propagator::B3::Multi.rack_extractor]
c.http_injectors = [OpenTelemetry::Propagator::B3::Single.rack_injector]
c.text_map_extractors = [OpenTelemetry::Propagator::B3::Single.text_map_extractor, OpenTelemetry::Propagator::B3::Multi.text_map_extractor]
c.text_map_injectors = [OpenTelemetry::Propagator::B3::Single.text_map_injector]
end
I’m working on adding OpenTelemetry to core puppet and was wondering if I need to do anything special since it can be
- a library in a MRI application
- a library in an application run on a JVM with JRuby that is made up of ruby, Java, and Cojure code (its launched like other Java apps)
- executed via being installed as a gem
1 reply
Most OTel SIGs are canceling meetings tomorrow because they conflict with the OpenTelemetry Community Day virtual conference: https://events.linuxfoundation.org/open-telemetry-community-day/. Should we do the same? I’ll be attending portions of the conference. Is anyone else planning to go?
4 replies
if anyone knows anyone, we're looking for a sr hire here and r actively recruiting , feel free to DM me if you know anyone https://www.datadoghq.com/careers/detail/?gh_jid=2147488
:wave: @ericmustin and @mwear open-telemetry/opentelemetry-ruby#341 is ready to review. I'm good with it and @robertlaurin is good with it, but it'd be good to get one more :+1: before we merge. Once this is in, I'll try to cut a new release tomorrow.
I pushed out a new release, but the release process seems to have failed after releasing the SDK and API. I think I'm messing things up by being over-eager about deleting the branch -- I think I should be leaving that to the release scripts. I kinda don't want to ping you on a holiday, Daniel, but if you happen to see this and can help unstick the release, I'd really appreciate it.
8 replies
i dont think i'll be able to attend the mtg today, I'm being tapped on the shoulder to do my best Lyle Lanley impression and do product demos for my employer at aws:reinvent :musical_note: I've sold monorails to Brockway, Ogdenville, and North Haverbrook, and, by gum, it put them on the map! :musical_note:
I sat in on the versioning meeting today. It sounds like a direction we may be going is to separate the various parts in the api (and maybe also the sdk) to separate packages, so that we can release unstable signals (like metrics) and not need to make semver major updates to stable signals (like trace). If we go that way, we might need to break the various subpackages out into separate gems, and then repurpose api to be "api-all-but-just-the-stable-ones"
1 reply
lemme work on making a repro
I think i fixed the problem. I’ll cut a new release today - maybe you can try that? IIRC, though, Robert installed it in a couple of example apps yesterday. We had issues with the Rails and “all” instrumentation gems - both fixed, but not yet released - i don’t think the sdk was a problem, though.
8 replies
Hey folks. There was an action item that came out of the maintainers meeting. We have been asked to:
- Read the Versioning and stability OTEP: open-telemetry/oteps#143
- Write up a draft (as a google doc) on how we will handle versioning and package releases. There are some details under the "What is in a language specific draft" section of the maintainers meeting notes: https://docs.google.com/document/d/1-tGqaT9fn9Bpuy4roLS1YDfLzIWHodUFfyrBlyJVz9g/edit#
- Post a link to the draft on this tracking issue: open-telemetry/opentelemetry-specification#1267
Perhaps we can try to read through the OTEP before our SIG meeting tomorrow and make some time to discuss how we'd like to handle things and what to put in our draft.
A starting point for Ruby's language-specific versioning: https://docs.google.com/document/d/1lz52CAtWS2k-SvbmDh-V6LgAj36h1bXcrwalIOM8VP8/edit
FYI, if anyone has any feedback for the Versioning and Stability OTEP it would be good to get it in soon: open-telemetry/oteps#143.
The Jaeger propagator should support baggage, but the baggage in API is intentionally no-op. Is it okay to just refer to the SDK baggage from within tests? I guess the more correct way to do that is to mock the baggage interface somehow but I'm wondering if that's necessary
1 reply
I have been struggling over many versions of otel-ruby to get some traces. I gave it another try recently now that
rails instrumentation and b3 propagation has hit (stuck with B3 as long as Istio/Envoy is using it). I see now that I get a Rack root span when I do a use_all. After that, all the redis spans get nested drive the stack too deep and everything is an error. Is this the result of open-telemetry/opentelemetry-ruby#352
3 replies
IMPORTANT: As discussed in the meeting this week, the default branch name has now been changed from
master to main. Existing pull requests, the CI and release configurations, and documentation links that point back at the GitHub tree, have all been updated accordingly. When you visit the repo on GitHub, it will show a popup with instructions for updating your local clone. Additionally, if you have a fork of the repo, you may consider updating its default branch name as well to avoid confusion. Feel free to ping me if you have problems/questions.
Hello,
I've just installed new version of the OT-Ruby instrumentation (0.13.0) and I'm testing it on https://github.com/JetBrains/sample_rails_app. I have few things:
I've just installed new version of the OT-Ruby instrumentation (0.13.0) and I'm testing it on https://github.com/JetBrains/sample_rails_app. I have few things:
- When I use OTLP exporter with
SimpleSpanProcessorI'm getting an error during span exportOpenTelemetry error: unexpected error in span.on_finish - undefined methodexport' for #<Hash:0x00007f954acc1a50>` - Looks like setting service name by environment variable
export OTEL_RESOURCE_ATTRIBUTES=service.name=testdoesn't work - I'm gettingunknown_service. (BatchSpanProcessor) - When I run example app from https://github.com/open-telemetry/opentelemetry-ruby/blob/opentelemetry-sdk/v0.13.0/instrumentation/rails/example/trace_request_demonstration.ru I expect to see
otlp.instrumentation.library.nametag with valueOpenTelemetry::Instrumentation::Railsbut it's reportingOpenTelemetry::Instrumentation::Rack. I'm not an Ruby expert but I would like to know if Rails instrumentation is really working :)
5 replies
I'm having a problem in our CI where I'm seeing
Could not find gem 'opentelemetry-api (~> 0.13.0)', which is required by gem
'opentelemetry-propagator-jaeger', in any of the relevant sources: source at../../api`` Any thoughts on what I can look into? This is a pattern used in many of the Gemfiles in the repo and I can't reproduce this issue locally
2 replies
:wave: Hi all - I had a question. I need to redact potentially sensitive data from exceptions recorded on spans, but for a variety of reasons cannot run a collector at this time - the scrubbing needs to be done in-process. I was wondering if anyone had any advice on the best approach to do this?
My ideal case would be that our developers can continue to call span.record_exception(error_object), and we scrub attributes before any exporter gets involved. But I'm not sure where I should hook into the pipeline:
- Should this be a custom SpanProcessor?
- Should this be a custom SpanExporter?
- Should I just monkey-patch
record_exception? :laughing:
Any advice you have is appreciated :)
7 replies
also when using the open telemetry with puma workers it doesn't see to find the context to push the metrics to agent.
Just made a README instrumentation issue: open-telemetry/opentelemetry-ruby#609 though I don't think I can assign myself
1 reply
Other SIGs seem to be moving to GitHub Discussions for ... discussions: https://github.com/open-telemetry/community#special-interest-groups while the "front door" for OTel is moving from Gitter to the CNCF Slack. Can/should we close this Gitter channel in favour of GitHub Discussions?
4 replies