Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
    Kranti Vikram A
    2 replies
    @here hi, Is there a way to marshal and unmarshal otlp ResourceSpans in contrib repo ? I am unable to access package otlptrace as its part of opentelemetry collector internal package. Please, suggest ?
    1 reply
    David Gildeh

    Hi everyone - got stuck on this issue for a few days now so hoping someone here can help! I'm running my Python Flask app on Google Cloud Run, and looking at traces I can't see the Request Trace ID sent to the Flask app propagated to the Trace context after following the example here:

    I'm expecting the traceId to match the request's traceId, but in Google Trace its showing a new traceId for my traces. What am I doing wrong? See my code below:

    from opentelemetry import trace
    from opentelemetry.exporter.cloud_trace import CloudTraceSpanExporter
    from opentelemetry.sdk.trace import TracerProvider
    from opentelemetry.sdk.trace.export import SimpleExportSpanProcessor
    from opentelemetry.propagators import set_global_httptextformat
    from import CloudTraceFormatPropagator
    from opentelemetry.instrumentation.flask import FlaskInstrumentor
    def initialize_tracer(app:Flask):
        # Setup tracer
        # Using the X-Cloud-Trace-Context header
    Hi, is there a documentation with a little more details than about tail sampling ? I did not get how use to it. For example, how I must configure it to have traces of all responses with status code 500 and a sampling of 10% of other requests ?
    3 replies
    Ankit Nayan
    I could not find any references for this in the codebase. Can anyone clear things here?
    4 replies
    Hi, do you have OpenTelemetry auto instrumentation for Maven projects? I cannot find it and I urgently need that. Thank you
    7 replies
    Hi, we're looking at running stand-alone collectors (tracing) for a multi-tenant environment. The idea would be a pipeline for each tennent that routes their traces to their dedicated jaeger back-end based on the incoming port (so each receiver maps to 1 exporter). In this setup, is there a recommended max number of pipelines per-instance of the collector?
    2 replies
    Ankit Nayan
    any docs/examples on how to use datadog auto instrumentation libraries?
    1 reply
    Sergey Kanzhelev
    Maintainer, approvers, and other active members. If you are interested to mentor people or have idea of the project, please comment: open-telemetry/community#468
    Is there plans to support writing metrics with HTTP/JSON using the OpenTelemetry Receiver in the OpenTelemetry Collector?
    9 replies
    Hello All, I was looking for open source Telemetry solutions and found page on github, Can someone share any link to understand the project at high level
    2 replies
    my usecase is getting information about all the network flows from a VM/Container/server e.g. source IP/DestinationIP/Portno. etc
    Daniel Norman

    Hey there,
    I'm new to open telemetry and have extensive experience with Prometheus.

    If I have a smaller Node.js application that I want to monitor, what are some of the hosted alternatives to Prometheus that support open telemetry?

    4 replies
    Hi All, I have a .NET core app that I'd like to export telemetry data to Azure Monitor/Application Insights. I could use the ApplicationInsights SDK but it seems more sensible to use OpenTelemetry as then we're not restricted to AppInsights. However, I can't find any good information on how to get C# data into Azure monitor. Other languages seem to have bespoke exporters but I am surprised there doesn't seem to be one for C#. A post in this room suggest there is an Azure Monitor exporter but the link no longer works and looking through the repo, it appears to be missing.
    11 replies
    Hello All,
    this is related to performance of opentelemetry-collector. I have deployed opentelemtry-collector in k8s. In my pipeline, I have otlp http receiver, attribute processor (adding one tag) and logging exporter. I have allotted 1cpu and 4gb memory.
    With this setup, at around 1250 request per second (1 span in each request), cpu utilization is well over 80%. With 3 spans in each request, cpu utilization was almost 100%. Each span has 10 attributes.
    Number seems far too less compared to the numbers mentioned in
    Appreciate any help/pointers on how to improve this performance.
    2 replies
    Michiel De Backker
    Hi everyone, I was wondering if there is a way to create metrics from spans? It feels like we're currently often doing double work: creating a span for tracing and creating a metric for dashboarding and alerting. I was wondering of there is an easy way (maybe inside the collector?) so we can automatically turn specific spans into metrics so we don't have to instrument twice?
    5 replies
    Cameron Moberg


    I'm trying to find consistent and explicit guidance on metric naming. Specifically I'm trying to figure out if slashes are supposed to be allowed or not in otel Java.

    Can someone point me to the correct documentation please?

    1 reply
    Tamas Eppel
    I receive the following exception in the JS library - using Google Cloud exporter
    (node:8) UnhandledPromiseRejectionWarning: TypeError: Cannot read property 'cloud.provider' of undefined
        at getTypeAndMappings (/node_modules/@google-cloud/opentelemetry-cloud-monitoring-exporter/build/src/transform.js:132:45)
    this tries to read
    I instantiate the exporter like so
    new MetricExporter({
                      prefix: metricsPrefix
    do I miss some config value?
    Evgeny Yakimov
    Hello, This may be one for the specification chat, but I'm wondering whether there has been any discussion/proposals about proxy collection/agents. For example in the context of client-side browser application, I would like to have spans collected on the client-side, but rather than send those traces directly to a collector/agent, I may want to handle them in my own front-end application server, for example, to enrich them with session/user data, or perhaps just to validate their authenticity.
    3 replies
    Hi, I'm new to Opentelemetry and I've been looking for a complete step by step documentation for an opensource platform that is written in java but I couldn't find anything helpful yet! can anyone introduce a helpful source for this purpose?
    Dhi Aurrahman
    For a new C++ project, to add tracing capability, are you suggesting on adopting OpenTracing/OpenCensus or using the currently WIP OpenTelemetry CPP SDK?
    Justin Pye
    How do I correlate js web traces with backend traces?
    2 replies
    Divyanshu Srivastava

    Hello Everyone,

    Someone please help me in this:-

    I was stuck and did not find any solution for that.

    6 replies
    Vinicius Vieira dos Santos
    Hello everyone, I'm currently using elastic apm with my java applications and I'm trying to migrate from their implementation of trace and spans to opentelemetry, however when I do that my spans stop being captured, I think I'm missing something in relation to "instrumentation -library-name ", I tried to put the name of my package and my application, but even then the span is lost.
    Would anyone have a working example of a Java application for me to take a look at?
    1 reply
    Vamsee Lakamsani
    Hello, the go OT collector examples appear to be relying on relative paths in the repo. How can we run them standalone/outside the repo by just relying on published go modules? For example, I tried building the basic example by removing the replace block but got an error.
    ➜  basic git:(master) ✗ cat go.mod
    go 1.14
    require ( v0.11.0 v0.11.0
    ➜  basic git:(master) ✗ go build
    go: finding module for package
    main.go:21:2: module found (v0.11.0), but does not contain package
    6 replies

    hello I am trying to exclude some spans based on the attribute http.url. I wan't to exclude all spans containing /actuator/health in path. I tried with this configuration :

              match_type: regexp
              span_names: ["^(.*)/actuator/health$"]
              separator: "::"
              from_attributes: [http.url]

    In the UI, I see that spans arre correctly renamed with http.url attribute but not filtered.

    3 replies
    Graham Lea
    Hi all. Relatively simple question. Is it part of the spec that all OT implementations should support multiple exporters? I can see it's possible in the Java & JS implementations, and in the Collector, but I couldn't find anywhere in the spec (using GH search) saying that this is a requirement of all implementations.
    Martin Foot
    Hello! I've got a Java process running as a Kubernetes Job. I've just followed a tutorial and have OpenTelemetry spans being published to Jaeger which I can see in the UI. I'd also like to push specific OpenMetrics metrics to Prometheus. Since it's a Job, I can't really have it listening waiting for Prometheus to scrape it, because it'll die after it's finished (or waste cluster resources). I can see Prometheus' Java client has a PushGateway class for pushing metrics, but I don't see an exporter that uses it ( Am I looking in the wrong place, or is the only exporter currently supported one that requires http polling?
    3 replies
    Aditya Prajapati
    Hey all,
    I'm a senior year cse undergrad interested in contributing to the project "OpenTelemetry to FluentBit exporter" under the q3-q4 of community bridge. I'm new to the community and am trying to learn more about the project by going through the specification, the next steps being setting up the project locally and trying it out. It would be really awesome if anyone could tell me if I'm going in the right direction and what else I should be doing :)
    I also wanted to know if there are any other eligibility criteria to be met other than the ones stated on the community bridge website, and the overall process to do so.
    P.S: While going through the logs specification, I was a bit confused by the current status of the project, so it'd be great if anyone could point me to where I can read more about it.
    P.P.S: Should I use gitter for future communication or is there any other alternative that I should stick to?
    3 replies
    Justin Pye
    Am I correct in saying that opentelemetry-go does not support communicating to the Collector over HTTP?
    Anthony Mirabella
    Correct, open-telemetry/opentelemetry-go#1122 exists to track adding support for HTTP/JSON and HTTP/Protobuf
    @IainAdamsLabs did you ever find a solution for .net core? I'd like to do the same and have had no luck getting either newrelic or azure working with 0.6.0-beta.1. New relic has an exporter but data does not show up on other end. Does azure monitor even have an exporter? I really need a solution in near future.
    2 replies
    Juraci Paixão Kröhling
    @mtwo , @SergeyKanzhelev , are you able to add me as a collaborator to the OpenTelemetry publication on Medium? I have a blog post that I would like to send there for consideration
    2 replies
    my username there is jpkroehling
    Graham Lea
    Is the ability to add multiple exporters part of the OTel spec, or just a bonus feature in some of the implementations?
    andres descalzo

    I'm trying to implement in my env k8s the "opentelemetry-collector-contrib", I'm using this
    and I replaced the image by "otel/opentelemetry-collector-contrib-dev", but when it was deployed this error happened

    Error: failed to create containerd task: OCI runtime create failed: container_linux.go:345: starting container process caused "exec: \"/otelcol\": stat /otelcol: no such file or directory": unknown

    When I use the original image "otel/opentelemetry-collector-contrib-dev" it works ok, but I need to export to apm-elastic.

    2 replies
    Hello a few days ago I found a repo that you could add a behavior to a WCF that wraps WCF calls with opentelemetry instrumentation but can't find it anymore . It was not that popular but it would help me a lot if someone knows about it
    1 reply
    Adam Flinton
    Hello, I am looking to do end to end telemetry on a distributed system preferably using the w3c trace context. I was wondering if open telemetry had javascript libs which could be used from in the browser so the entire set of hops can be traced.
    1 reply
    Reginald Davis
    Hey everyone, I've been attempting to instrument a Angular 8 app with tracing. Can anyone point me to an example to help out?
    4 replies
    Hi everyone, I'm using the Jaeger OTEL collector and would like to set it up to sample 10% of every traces and 100% of every errors (e.g. http.status_code >= 400). Right now I've set up a couple pipelines, one to handle errors with tail_sampling and another one to handle successes with tail_sampling + 10% proba. They both use the same receivers and exporters, only the processors differ. It works but this seems a bit hacky, is there a better way to handle this?
    10 replies
    David Prenninger
    Hey there!
    I have a question regarding JMH benchmarking with the auto instrumenter. I never really worked with JMH before, but figured the metrics to be very handy.
    What exactly happens in the scenarios ClassRetransformingBenchmark.WithAgent.testTracedRetransform and testUntracedRetransform? And what is the difference to HttpBenchmark.WithAgent.testTracedRetransform?
    Hope these questions are not to specific O.o
    This is a repost from the opentelemetry-java-instrumentation channel.
    Grzegorz Polek
    Hi all, where I can find some additional info regarding OpenMetrics support in OpenTelemetry? I've found this OpenObservability/OpenMetrics#137 reply in OpenMetrics repo and I would like to know, what is the best approach to OpenTelemetry and OpenMetrics.
    Currently, we are extensively using Prometheus for all metrics produced by our services, and when we introduce OpenTelemetry, we would have two separate ways of exposing metrics from the service (via OpenTelemetry, and via scrapping endpoint with Prometheus).
    2 replies
    bbigras Is still a thing? Got a page not found.
    Severin Neumann

    Hello everyone,

    as a newcomer to OpenTelemetry I am currently figuring out how to get started and where I can contribute to the project.
    I love doing new things, and I enjoy it even more when I can learn with and from other people: it goes so much faster and makes so much more fun.
    For this, I’d like to set up a virtual event within the Hacktoberfest (see also open-telemetry/community#474), but I need help from existing contributors and other people that are new!
    So, if you are also looking for an opportunity to learn and have fun together, let me know!

    5 replies
    Endre Karlson
    Hi people! Is it so that I can use opentelemetry-collector as a in-cluster replacement for prometheus that just forwards then to a centralized prometheus remote-write enabled instance?
    Marius Jokubauskas
    Hi, I have thought I'll make a quick look at OT, and installed docker example form and java auto instrumentation agent, but seems that there is no connection between these two parts: javaaagent complaining "$AnnotatedConnectException: Connection refused: no further information: localhost/0:0:0:0:0:0:0:1:55678"; I have changed no options, everything is default