feat: 增强 OpenTelemetry 提供者配置,添加连接安全、采样器和批处理选项

This commit is contained in:
Rogee
2025-09-11 18:23:14 +08:00
parent 9662d7d718
commit 9e7ce1b7d3
4 changed files with 500 additions and 262 deletions

View File

@@ -458,3 +458,56 @@ import (
// 仅需 import 即可,无额外 ServerOption // 仅需 import 即可,无额外 ServerOption
``` ```
## OpenTelemetry 集成(推荐)
使用 StatsHandler推荐不与拦截器同时使用避免重复埋点
```go
import (
otelgrpc "go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc"
)
// 基本接入:使用全局 Tracer/Meter由 OTEL Provider 初始化)
handler := otelgrpc.NewServerHandler(
otelgrpc.WithTraceEvents(), // 在 span 中记录消息事件
)
pgrpc.UseOptions(grpc.StatsHandler(handler))
// 忽略某些方法(如健康检查),避免噪声:
handler = otelgrpc.NewServerHandler(
otelgrpc.WithFilter(func(ctx context.Context, fullMethod string) bool {
return fullMethod != "/grpc.health.v1.Health/Check"
}),
)
pgrpc.UseOptions(grpc.StatsHandler(handler))
```
使用拦截器版本(如你更偏好 Interceptor 方案;与 StatsHandler 二选一):
```go
import (
otelgrpc "go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc"
)
pgrpc.UseUnaryInterceptors(otelgrpc.UnaryServerInterceptor())
pgrpc.UseStreamInterceptors(otelgrpc.StreamServerInterceptor())
```
> 注意:不要同时启用 StatsHandler 和拦截器,否则会重复生成 span/metrics。
## OpenTracingJaeger集成
当使用 Tracing ProviderJaeger + OpenTracing可使用 opentracing 的 gRPC 拦截器:
```go
import (
opentracing "github.com/opentracing/opentracing-go"
otgrpc "github.com/grpc-ecosystem/grpc-opentracing/go/otgrpc"
)
pgrpc.UseUnaryInterceptors(otgrpc.OpenTracingServerInterceptor(opentracing.GlobalTracer()))
pgrpc.UseStreamInterceptors(otgrpc.OpenTracingStreamServerInterceptor(opentracing.GlobalTracer()))
```
> 与 OTEL 方案互斥:如果已启用 OTEL请不要再开启 OpenTracing 拦截器,以免重复埋点。

View File

@@ -0,0 +1,121 @@
OpenTelemetry Provider (OTLP Traces + Metrics)
该 Provider 基于 OpenTelemetry Go SDK初始化全局 Tracer 与 Meter支持 OTLP(gRPC/HTTP) 导出,并收集运行时指标。
配置config.toml
```
[OTEL]
ServiceName = "my-service"
Version = "1.0.0"
Env = "dev"
# 导出端点(二选一)
EndpointGRPC = "otel-collector:4317"
EndpointHTTP = "otel-collector:4318"
# 认证(可选)
Token = "Bearer <your-token>" # 也可只填纯 tokenProvider 会自动补齐 Bearer 前缀
# 安全(可选)
InsecureGRPC = true # gRPC 导出是否使用 insecure
InsecureHTTP = true # HTTP 导出是否使用 insecure
# 采样(可选)
Sampler = "always" # always|ratio
SamplerRatio = 0.1 # Sampler=ratio 时生效0..1
# 批处理(可选,毫秒)
BatchTimeoutMs = 5000
ExportTimeoutMs = 10000
MaxQueueSize = 2048
MaxExportBatchSize = 512
# 指标(可选,毫秒)
MetricReaderIntervalMs = 10000 # 指标导出周期
RuntimeReadMemStatsIntervalMs = 5000 # 运行时指标读取周期
```
启用
```
import "test/providers/otel"
func providers() container.Providers {
return container.Providers{
otel.DefaultProvider(),
}
}
```
使用
- Traces: 通过 `go.opentelemetry.io/otel` 获取全局 Tracer或使用仓库提供的 `providers/otel/funcs.go` 包装。
```
ctx, span := otel.Tracer("my-service").Start(ctx, "my-op")
// ...
span.End()
```
- Metrics: 通过 `otel.Meter("my-service")` 创建仪表,或使用 `providers/otel/funcs.go` 的便捷函数。
与 Tracing Provider 的区别与场景建议
- Tracing ProviderJaeger + OpenTracing只做链路适合已有 OpenTracing 项目;
- OTEL ProviderOpenTelemetry统一 Traces+Metrics对接 OTLP 生态,适合新项目或希望统一可观测性;
- 可先混用:保留 Jaeger 链路,同时启用 OTEL 运行时指标,逐步迁移。
快速启动(本地 Collector
最小化 docker-compose
```
services:
otel-collector:
image: otel/opentelemetry-collector:0.104.0
command: ["--config=/etc/otelcol-config.yml"]
volumes:
- ./otelcol-config.yml:/etc/otelcol-config.yml:ro
ports:
- "4317:4317" # OTLP gRPC
- "4318:4318" # OTLP HTTP
```
示例 otelcol-config.yml
```
receivers:
otlp:
protocols:
grpc:
http:
exporters:
debug:
verbosity: detailed
processors:
batch:
service:
pipelines:
traces:
receivers: [otlp]
processors: [batch]
exporters: [debug]
metrics:
receivers: [otlp]
processors: [batch]
exporters: [debug]
```
应用端:
```
[OTEL]
EndpointGRPC = "127.0.0.1:4317"
InsecureGRPC = true
```
故障与降级
- Collector/网络异常OTEL SDK 异步批处理,不阻塞业务;可能丢点/丢指标;
- 启动失败:初始化报错会阻止启动;如需“不可达也不影响启动”,可加开关降级为 no-op可按需补充

View File

@@ -28,6 +28,25 @@ type Config struct {
EndpointGRPC string EndpointGRPC string
EndpointHTTP string EndpointHTTP string
Token string Token string
// Connection security
InsecureGRPC bool // if true, use grpc insecure for OTLP gRPC
InsecureHTTP bool // if true, use http insecure for OTLP HTTP
// Tracing sampler
// Sampler: "always" (default) or "ratio"
Sampler string
SamplerRatio float64 // used when Sampler == "ratio"; 0..1
// Tracing batcher options (milliseconds)
BatchTimeoutMs uint
ExportTimeoutMs uint
MaxQueueSize int
MaxExportBatchSize int
// Metrics options (milliseconds)
MetricReaderIntervalMs uint // export interval for PeriodicReader
RuntimeReadMemStatsIntervalMs uint // runtime metrics min read interval
} }
func (c *Config) format() { func (c *Config) format() {

View File

@@ -26,6 +26,17 @@ import (
"google.golang.org/grpc/encoding/gzip" "google.golang.org/grpc/encoding/gzip"
) )
// formatAuth formats token into an Authorization header value.
func formatAuth(token string) string {
if token == "" {
return ""
}
if len(token) > 7 && (token[:7] == "Bearer " || token[:7] == "bearer ") {
return token
}
return "Bearer " + token
}
func Provide(opts ...opt.Option) error { func Provide(opts ...opt.Option) error {
o := opt.New(opts...) o := opt.New(opts...)
var config Config var config Config
@@ -82,7 +93,7 @@ func (o *builder) initResource(ctx context.Context) (err error) {
semconv.HostNameKey.String(hostName), // semconv.HostNameKey.String(hostName), //
), ),
) )
return return err
} }
func (o *builder) initMeterProvider(ctx context.Context) (err error) { func (o *builder) initMeterProvider(ctx context.Context) (err error) {
@@ -92,9 +103,8 @@ func (o *builder) initMeterProvider(ctx context.Context) (err error) {
otlpmetricgrpc.WithCompressor(gzip.Name), otlpmetricgrpc.WithCompressor(gzip.Name),
} }
if o.config.Token != "" { if h := formatAuth(o.config.Token); h != "" {
headers := map[string]string{"Authentication": o.config.Token} opts = append(opts, otlpmetricgrpc.WithHeaders(map[string]string{"Authorization": h}))
opts = append(opts, otlpmetricgrpc.WithHeaders(headers))
} }
exporter, err := otlpmetricgrpc.New(ctx, opts...) exporter, err := otlpmetricgrpc.New(ctx, opts...)
@@ -109,9 +119,11 @@ func (o *builder) initMeterProvider(ctx context.Context) (err error) {
otlpmetrichttp.WithEndpoint(o.config.EndpointHTTP), otlpmetrichttp.WithEndpoint(o.config.EndpointHTTP),
otlpmetrichttp.WithCompression(1), otlpmetrichttp.WithCompression(1),
} }
if o.config.InsecureHTTP {
if o.config.Token != "" { opts = append(opts, otlpmetrichttp.WithInsecure())
opts = append(opts, otlpmetrichttp.WithURLPath(o.config.Token)) }
if h := formatAuth(o.config.Token); h != "" {
opts = append(opts, otlpmetrichttp.WithHeaders(map[string]string{"Authorization": h}))
} }
exporter, err := otlpmetrichttp.New(ctx, opts...) exporter, err := otlpmetrichttp.New(ctx, opts...)
@@ -129,18 +141,27 @@ func (o *builder) initMeterProvider(ctx context.Context) (err error) {
} }
if err != nil { if err != nil {
return return err
} }
// periodic reader with optional custom interval
var readerOpts []sdkmetric.PeriodicReaderOption
if o.config.MetricReaderIntervalMs > 0 {
readerOpts = append(readerOpts, sdkmetric.WithInterval(time.Duration(o.config.MetricReaderIntervalMs)*time.Millisecond))
}
meterProvider := sdkmetric.NewMeterProvider( meterProvider := sdkmetric.NewMeterProvider(
sdkmetric.WithReader( sdkmetric.WithReader(
sdkmetric.NewPeriodicReader(exporter), sdkmetric.NewPeriodicReader(exporter, readerOpts...),
), ),
sdkmetric.WithResource(o.resource), sdkmetric.WithResource(o.resource),
) )
otel.SetMeterProvider(meterProvider) otel.SetMeterProvider(meterProvider)
err = runtime.Start(runtime.WithMinimumReadMemStatsInterval(time.Second * 5)) interval := 5 * time.Second
if o.config.RuntimeReadMemStatsIntervalMs > 0 {
interval = time.Duration(o.config.RuntimeReadMemStatsIntervalMs) * time.Millisecond
}
err = runtime.Start(runtime.WithMinimumReadMemStatsInterval(interval))
if err != nil { if err != nil {
return errors.Wrapf(err, "Failed to start runtime metrics") return errors.Wrapf(err, "Failed to start runtime metrics")
} }
@@ -151,7 +172,7 @@ func (o *builder) initMeterProvider(ctx context.Context) (err error) {
} }
}) })
return return err
} }
func (o *builder) initTracerProvider(ctx context.Context) error { func (o *builder) initTracerProvider(ctx context.Context) error {
@@ -159,15 +180,13 @@ func (o *builder) initTracerProvider(ctx context.Context) error {
opts := []otlptracegrpc.Option{ opts := []otlptracegrpc.Option{
otlptracegrpc.WithCompressor(gzip.Name), otlptracegrpc.WithCompressor(gzip.Name),
otlptracegrpc.WithEndpoint(o.config.EndpointGRPC), otlptracegrpc.WithEndpoint(o.config.EndpointGRPC),
otlptracegrpc.WithInsecure(), // }
if o.config.InsecureGRPC {
opts = append(opts, otlptracegrpc.WithInsecure())
} }
if o.config.Token != "" { if h := formatAuth(o.config.Token); h != "" {
headers := map[string]string{ opts = append(opts, otlptracegrpc.WithHeaders(map[string]string{"Authorization": h}))
"Authentication": o.config.Token,
"authorization": o.config.Token, //
}
opts = append(opts, otlptracegrpc.WithHeaders(headers))
} }
log.Debugf("Creating GRPC trace exporter with endpoint: %s", o.config.EndpointGRPC) log.Debugf("Creating GRPC trace exporter with endpoint: %s", o.config.EndpointGRPC)
@@ -189,17 +208,15 @@ func (o *builder) initTracerProvider(ctx context.Context) error {
exporterHttpFunc := func(ctx context.Context) (*otlptrace.Exporter, error) { exporterHttpFunc := func(ctx context.Context) (*otlptrace.Exporter, error) {
opts := []otlptracehttp.Option{ opts := []otlptracehttp.Option{
otlptracehttp.WithInsecure(),
otlptracehttp.WithCompression(1), otlptracehttp.WithCompression(1),
otlptracehttp.WithEndpoint(o.config.EndpointHTTP), otlptracehttp.WithEndpoint(o.config.EndpointHTTP),
} }
if o.config.InsecureHTTP {
opts = append(opts, otlptracehttp.WithInsecure())
}
if o.config.Token != "" { if h := formatAuth(o.config.Token); h != "" {
opts = append(opts, opts = append(opts, otlptracehttp.WithHeaders(map[string]string{"Authorization": h}))
otlptracehttp.WithHeaders(map[string]string{
"Authentication": o.config.Token,
}),
)
} }
log.Debugf("Creating HTTP trace exporter with endpoint: %s", o.config.EndpointHTTP) log.Debugf("Creating HTTP trace exporter with endpoint: %s", o.config.EndpointHTTP)
@@ -225,10 +242,38 @@ func (o *builder) initTracerProvider(ctx context.Context) error {
return err return err
} }
// Sampler
sampler := sdktrace.AlwaysSample()
if o.config.Sampler == "ratio" {
ratio := o.config.SamplerRatio
if ratio <= 0 {
ratio = 0
}
if ratio > 1 {
ratio = 1
}
sampler = sdktrace.ParentBased(sdktrace.TraceIDRatioBased(ratio))
}
// Batcher options
var batchOpts []sdktrace.BatchSpanProcessorOption
if o.config.BatchTimeoutMs > 0 {
batchOpts = append(batchOpts, sdktrace.WithBatchTimeout(time.Duration(o.config.BatchTimeoutMs)*time.Millisecond))
}
if o.config.ExportTimeoutMs > 0 {
batchOpts = append(batchOpts, sdktrace.WithExportTimeout(time.Duration(o.config.ExportTimeoutMs)*time.Millisecond))
}
if o.config.MaxQueueSize > 0 {
batchOpts = append(batchOpts, sdktrace.WithMaxQueueSize(o.config.MaxQueueSize))
}
if o.config.MaxExportBatchSize > 0 {
batchOpts = append(batchOpts, sdktrace.WithMaxExportBatchSize(o.config.MaxExportBatchSize))
}
traceProvider := sdktrace.NewTracerProvider( traceProvider := sdktrace.NewTracerProvider(
sdktrace.WithSampler(sdktrace.AlwaysSample()), sdktrace.WithSampler(sampler),
sdktrace.WithResource(o.resource), sdktrace.WithResource(o.resource),
sdktrace.WithBatcher(exporter), sdktrace.WithBatcher(exporter, batchOpts...),
) )
container.AddCloseAble(func() { container.AddCloseAble(func() {
log.Error("shut down") log.Error("shut down")