将OTLP追踪数据导出至Jaeger时遇服务未就绪错误求助

尝试将OTLP追踪数据导出至Jaeger时遇到如下错误：

OpenTelemetry trace error occurred. Exporter otlp encountered the following error(s): the grpc server returns error (Unknown error): , detailed error message: Service was not ready: transport error

推测可能是tokio与tonic的交互问题，但无法控制客户端创建。尝试使用grpc方式也失败，http方式缺乏文档说明。

可复现示例 #

Rust代码 #

use opentelemetry::trace::Tracer;
use opentelemetry_otlp::WithExportConfig;

#[tokio::main]
async fn main() {
    let otlp_exporter = opentelemetry_otlp::new_exporter().tonic().with_endpoint("http://jaeger:4317/");
    let tracer = opentelemetry_otlp::new_pipeline()
        .tracing()
        .with_exporter(otlp_exporter)
        .install_simple()
        .unwrap();

    tracer.in_span("xyz", |_| {});
}

Cargo配置 #

[package]
name = "sample"
version = "0.1.0"
edition = "2021"

[dependencies]
opentelemetry = "0.18.0"
opentelemetry-otlp = "0.11"
tokio = { version = "1.28", features = ["rt-multi-thread"] }

Docker Compose配置 #

version: '3.8'
services:
  app:
    image: mcr.microsoft.com/devcontainers/base:jammy
    volumes:
      - ../..:/workspaces:cached
    command: sleep infinity
    network_mode: service:jaeger
  jaeger:
    image: jaegertracing/all-in-one:1.44
    restart: unless-stopped

Jaeger运行在相邻容器中，DNS名称jaeger可解析。

以下是几个可行的排查和修复方向：

1. 强制启用Jaeger的OTLP GRPC端口 #

Jaeger all-in-one镜像默认开启4317端口，但可通过环境变量明确启用。修改Compose中的Jaeger服务：

jaeger:
  image: jaegertracing/all-in-one:1.44
  restart: unless-stopped
  environment:
    - COLLECTOR_OTLP_ENABLED=true

重启Jaeger容器后重新测试连接。

2. 切换为OTLP HTTP/JSON协议 #

如果GRPC方式存在兼容性问题，可尝试HTTP/JSON协议：
首先更新Cargo.toml，添加http特性：

opentelemetry-otlp = { version = "0.11", features = ["http"] }

然后修改代码使用HTTP exporter：

use opentelemetry::trace::Tracer;
use opentelemetry_otlp::WithExportConfig;

#[tokio::main]
async fn main() {
    let otlp_exporter = opentelemetry_otlp::new_exporter()
        .http()
        .with_endpoint("http://jaeger:4318/v1/traces"); // Jaeger OTLP HTTP默认端口为4318
    let tracer = opentelemetry_otlp::new_pipeline()
        .tracing()
        .with_exporter(otlp_exporter)
        .install_simple()
        .unwrap();

    tracer.in_span("xyz", |_| {});
}

3. 修复依赖版本兼容性 #

版本不匹配可能导致tokio与tonic交互异常，更新依赖到兼容版本：

[dependencies]
opentelemetry = "0.19.0"
opentelemetry-otlp = "0.12.0"
tokio = { version = "1.32", features = ["rt-multi-thread", "macros"] }
tonic = "0.9" # 显式指定版本，确保与opentelemetry-otlp兼容

更新后重新构建项目测试。

4. 验证容器网络连通性 #

在app容器内手动测试端口连通性：

# 进入app容器
docker exec -it <app_container_id> /bin/bash
# 测试GRPC端口
nc -zv jaeger 4317
# 测试HTTP端口
nc -zv jaeger 4318

若端口不通，检查Jaeger容器是否正常启动，或是否有网络限制。

5. 启用调试日志排查细节 #

添加日志依赖并启用调试输出，帮助定位问题：
首先更新Cargo.toml：

tracing = "0.1"
tracing-subscriber = { version = "0.3", features = ["env-filter"] }

然后修改代码：

use opentelemetry::trace::Tracer;
use opentelemetry_otlp::WithExportConfig;
use tracing_subscriber::EnvFilter;

#[tokio::main]
async fn main() {
    // 启用调试日志
    tracing_subscriber::fmt()
        .with_env_filter(EnvFilter::from_default_env())
        .init();

    let otlp_exporter = opentelemetry_otlp::new_exporter()
        .tonic()
        .with_endpoint("http://jaeger:4317/")
        .with_timeout(std::time::Duration::from_secs(10)); // 设置超时时间
    let tracer = opentelemetry_otlp::new_pipeline()
        .tracing()
        .with_exporter(otlp_exporter)
        .install_simple()
        .unwrap();

    tracer.in_span("xyz", |_| {});

    // 等待追踪数据导出完成
    opentelemetry::global::shutdown_tracer_provider();
}

运行时设置环境变量RUST_LOG=debug查看详细日志，获取连接过程中的具体错误信息。

内容的提问来源于stack exchange，提问作者Tom01098