Tracing: Standardize on otel tracing (grafana#75528)

contribute/backend/instrumentation.md

@@ -156,7 +156,7 @@ A distributed trace is data that tracks an application request as it flows throu
 
 ### Usage
 
-Grafana currently supports two tracing implementations, [OpenTelemetry](https://opentelemetry.io/) and [OpenTracing](https://opentracing.io/). OpenTracing is deprecated, but still supported until we remove it. The two different implementations implements the `Tracer` and `Span` interfaces, defined in the _pkg/infra/tracing_ package, which you can use to create traces and spans. To get a hold of a `Tracer` you would need to get it injected as dependency into your service, see [Services](services.md) for more details.
+Grafana uses [OpenTelemetry](https://opentelemetry.io/) for distributed tracing. There's an interface `Tracer` in the _pkg/infra/tracing_ package that implements the [OpenTelemetry Tracer interface](go.opentelemetry.io/otel/trace), which you can use to create traces and spans. To get a hold of a `Tracer` you would need to get it injected as dependency into your service, see [Services](services.md) for more details. For more information, see https://opentelemetry.io/docs/instrumentation/go/manual/.
 
 Example:
 
@@ -166,6 +166,7 @@ import (
 
    "github.com/grafana/grafana/pkg/infra/tracing"
    "go.opentelemetry.io/otel/attribute"
+   "go.opentelemetry.io/otel/trace"
 )
 
 type MyService struct {
@@ -179,36 +180,36 @@ func ProvideService(tracer tracing.Tracer) *MyService {
 }
 
 func (s *MyService) Hello(ctx context.Context, name string) (string, error) {
-   ctx, span := s.tracer.Start(ctx, "MyService.Hello")
+   ctx, span := s.tracer.Start(ctx, "MyService.Hello", trace.WithAttributes(
+      attribute.String("my_attribute", "val"),
+   ))
    // this make sure the span is marked as finished when this
    // method ends to allow the span to be flushed and sent to
    // storage backend.
    defer span.End()
 
    // Add some event to show Events usage
-   span.AddEvents(
-      []string{"message"},
-      []tracing.EventValue{
-         {Str: "checking name..."},
-      })
+   span.AddEvent("checking name...")
 
    if name == "" {
       err := fmt.Errorf("name cannot be empty")
 
+      // sets the span’s status to Error to make the span tracking
+      // a failed operation as an error span.
+      span.SetStatus(codes.Error, "failed to check name")
       // record err as an exception span event for this span
       span.RecordError(err)
       return "", err
    }
 
    // Add some other event to show Events usage
-   span.AddEvents(
-      []string{"message"},
-      []tracing.EventValue{
-         {Str: "name checked"},
-      })
+   span.AddEvent("name checked")
 
    // Add attribute to show Attributes usage
-   span.SetAttributes("my_service.name", name, attribute.Key("my_service.name").String(name))
+   span.SetAttributes(
+      attribute.String("my_service.name", name),
+      attribute.Int64("my_service.some_other", int64(1337)),
+   )
 
    return fmt.Sprintf("Hello %s", name), nil
 }
@@ -243,6 +244,22 @@ If span names, attribute or event values originates from user input they **shoul
 
 Be **careful** to not expose any sensitive information in span names, attribute or event values, e.g. secrets, credentials etc.
 
+### Span attributes
+
+Consider using `attributes.<Type>("<key>", <value>)` in favor of `attributes.Key("<key>").<Type>(<value>)` since it requires less characters and thereby reads easier.
+
+Example:
+
+```go
+attribute.String("datasource_name", proxy.ds.Name)
+// vs
+attribute.Key("datasource_name").String(proxy.ds.Name)
+
+attribute.Int64("org_id", proxy.ctx.SignedInUser.OrgID)
+// vs
+attribute.Key("org_id").Int64(proxy.ctx.SignedInUser.OrgID)
+```
+
 ### How to collect, visualize and query traces (and correlate logs with traces) locally
 
 #### 1. Start Jaeger
@@ -255,20 +272,11 @@ make devenv sources=jaeger
 
 To enable tracing in Grafana, you must set the address in your config.ini file
 
-opentelemetry tracing (recommended):
-
 ```ini
 [tracing.opentelemetry.jaeger]
 address = http://localhost:14268/api/traces
 ```
 
-opentracing tracing (deprecated/not recommended):
-
-```ini
-[tracing.jaeger]
-address = localhost:6831
-```
-
 #### 3. Search/browse collected logs and traces in Grafana Explore
 
 You need provisioned gdev-jaeger and gdev-loki datasources, see [developer dashboard and data sources](https://github.com/grafana/grafana/tree/main/devenv#developer-dashboards-and-data-sources) for setup instructions.

go.mod

@@ -121,7 +121,7 @@ require (
 	gopkg.in/mail.v2 v2.3.1 // @grafana/backend-platform
 	gopkg.in/yaml.v2 v2.4.0 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // @grafana/alerting-squad-backend
-	xorm.io/builder v0.3.6 // indirect; @grafana/backend-platform
+	xorm.io/builder v0.3.6 // @grafana/backend-platform
 	xorm.io/core v0.7.3 // @grafana/backend-platform
 	xorm.io/xorm v0.8.2 // @grafana/alerting-squad-backend
 )
@@ -174,7 +174,7 @@ require (
 	github.com/grpc-ecosystem/go-grpc-prometheus v1.2.1-0.20191002090509-6af20e3a5340 // indirect
 	github.com/hashicorp/errwrap v1.1.0 // indirect
 	github.com/hashicorp/go-msgpack v0.5.5 // indirect
-	github.com/hashicorp/go-multierror v1.1.1 // indirect; @grafana/grafana-as-code
+	github.com/hashicorp/go-multierror v1.1.1 // @grafana/grafana-as-code
 	github.com/hashicorp/go-sockaddr v1.0.2 // indirect
 	github.com/hashicorp/golang-lru v0.6.0 // indirect
 	github.com/hashicorp/yamux v0.1.1 // indirect

pkg/api/pluginproxy/ds_proxy.go

@@ -12,6 +12,7 @@ import (
 	"time"
 
 	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/trace"
 
 	"github.com/grafana/grafana/pkg/api/datasource"
 	"github.com/grafana/grafana/pkg/infra/httpclient"
@@ -142,10 +143,12 @@ func (proxy *DataSourceProxy) HandleRequest() {
 
 	proxy.ctx.Req = proxy.ctx.Req.WithContext(ctx)
 
-	span.SetAttributes("datasource_name", proxy.ds.Name, attribute.Key("datasource_name").String(proxy.ds.Name))
-	span.SetAttributes("datasource_type", proxy.ds.Type, attribute.Key("datasource_type").String(proxy.ds.Type))
-	span.SetAttributes("user", proxy.ctx.SignedInUser.Login, attribute.Key("user").String(proxy.ctx.SignedInUser.Login))
-	span.SetAttributes("org_id", proxy.ctx.SignedInUser.OrgID, attribute.Key("org_id").Int64(proxy.ctx.SignedInUser.OrgID))
+	span.SetAttributes(
+		attribute.String("datasource_name", proxy.ds.Name),
+		attribute.String("datasource_type", proxy.ds.Type),
+		attribute.String("user", proxy.ctx.SignedInUser.Login),
+		attribute.Int64("org_id", proxy.ctx.SignedInUser.OrgID),
+	)
 
 	proxy.addTraceFromHeaderValue(span, "X-Panel-Id", "panel_id")
 	proxy.addTraceFromHeaderValue(span, "X-Dashboard-Id", "dashboard_id")
@@ -155,11 +158,11 @@ func (proxy *DataSourceProxy) HandleRequest() {
 	reverseProxy.ServeHTTP(proxy.ctx.Resp, proxy.ctx.Req)
 }
 
-func (proxy *DataSourceProxy) addTraceFromHeaderValue(span tracing.Span, headerName string, tagName string) {
+func (proxy *DataSourceProxy) addTraceFromHeaderValue(span trace.Span, headerName string, tagName string) {
 	panelId := proxy.ctx.Req.Header.Get(headerName)
 	dashId, err := strconv.Atoi(panelId)
 	if err == nil {
-		span.SetAttributes(tagName, dashId, attribute.Key(tagName).Int(dashId))
+		span.SetAttributes(attribute.Int(tagName, dashId))
 	}
 }
 

pkg/api/pluginproxy/pluginproxy.go

@@ -109,8 +109,10 @@ func (proxy *PluginProxy) HandleRequest() {
 
 	proxy.ctx.Req = proxy.ctx.Req.WithContext(ctx)
 
-	span.SetAttributes("user", proxy.ctx.SignedInUser.Login, attribute.Key("user").String(proxy.ctx.SignedInUser.Login))
-	span.SetAttributes("org_id", proxy.ctx.SignedInUser.OrgID, attribute.Key("org_id").Int64(proxy.ctx.SignedInUser.OrgID))
+	span.SetAttributes(
+		attribute.String("user", proxy.ctx.SignedInUser.Login),
+		attribute.Int64("org_id", proxy.ctx.SignedInUser.OrgID),
+	)
 
 	proxy.tracer.Inject(ctx, proxy.ctx.Req.Header, span)
 

pkg/bus/bus.go

@@ -55,7 +55,7 @@ func (b *InProcBus) Publish(ctx context.Context, msg Msg) error {
 	_, span := b.tracer.Start(ctx, "bus - "+msgName)
 	defer span.End()
 
-	span.SetAttributes("msg", msgName, attribute.Key("msg").String(msgName))
+	span.SetAttributes(attribute.String("msg", msgName))
 
 	return nil
 }

pkg/expr/commands.go

@@ -70,7 +70,7 @@ func (gm *MathCommand) NeedsVars() []string {
 // failed to execute.
 func (gm *MathCommand) Execute(ctx context.Context, _ time.Time, vars mathexp.Vars, tracer tracing.Tracer) (mathexp.Results, error) {
 	_, span := tracer.Start(ctx, "SSE.ExecuteMath")
-	span.SetAttributes("expression", gm.RawExpression, attribute.Key("expression").String(gm.RawExpression))
+	span.SetAttributes(attribute.String("expression", gm.RawExpression))
 	defer span.End()
 	return gm.Expression.Execute(gm.refID, vars, tracer)
 }
@@ -163,7 +163,7 @@ func (gr *ReduceCommand) Execute(ctx context.Context, _ time.Time, vars mathexp.
 	_, span := tracer.Start(ctx, "SSE.ExecuteReduce")
 	defer span.End()
 
-	span.SetAttributes("reducer", gr.Reducer, attribute.Key("reducer").String(gr.Reducer))
+	span.SetAttributes(attribute.String("reducer", gr.Reducer))
 
 	newRes := mathexp.Results{}
 	for i, val := range vars[gr.VarToReduce].Values {

pkg/expr/dataplane.go

@@ -61,7 +61,7 @@ func shouldUseDataplane(frames data.Frames, logger log.Logger, disable bool) (dt
 func handleDataplaneFrames(ctx context.Context, tracer tracing.Tracer, t data.FrameType, frames data.Frames) (mathexp.Results, error) {
 	_, span := tracer.Start(ctx, "SSE.HandleDataPlaneData")
 	defer span.End()
-	span.SetAttributes("dataplane.type", t, attribute.Key("dataplane.type").String(string(t)))
+	span.SetAttributes(attribute.String("dataplane.type", string(t)))
 
 	switch t.Kind() {
 	case data.KindUnknown:

pkg/expr/graph.go

@@ -99,10 +99,10 @@ func (dp *DataPipeline) execute(c context.Context, now time.Time, s *Service) (m
 		}
 
 		c, span := s.tracer.Start(c, "SSE.ExecuteNode")
-		span.SetAttributes("node.refId", node.RefID(), attribute.Key("node.refId").String(node.RefID()))
+		span.SetAttributes(attribute.String("node.refId", node.RefID()))
 		if len(node.NeedsVars()) > 0 {
 			inputRefIDs := node.NeedsVars()
-			span.SetAttributes("node.inputRefIDs", inputRefIDs, attribute.Key("node.inputRefIDs").StringSlice(inputRefIDs))
+			span.SetAttributes(attribute.StringSlice("node.inputRefIDs", inputRefIDs))
 		}
 		defer span.End()
 

pkg/expr/nodes.go

@@ -10,12 +10,12 @@ import (
 	"github.com/grafana/grafana-plugin-sdk-go/backend"
 	"github.com/grafana/grafana-plugin-sdk-go/data"
 	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/codes"
 	"gonum.org/v1/gonum/graph/simple"
 
 	"github.com/grafana/grafana/pkg/expr/classic"
 	"github.com/grafana/grafana/pkg/expr/mathexp"
 	"github.com/grafana/grafana/pkg/infra/log"
-	"github.com/grafana/grafana/pkg/infra/tracing"
 	"github.com/grafana/grafana/pkg/services/datasources"
 	"github.com/grafana/grafana/pkg/services/featuremgmt"
 )
@@ -236,8 +236,10 @@ func executeDSNodesGrouped(ctx context.Context, now time.Time, vars mathexp.Vars
 				"datasourceVersion", firstNode.datasource.Version,
 			)
 
-			span.SetAttributes("datasource.type", firstNode.datasource.Type, attribute.Key("datasource.type").String(firstNode.datasource.Type))
-			span.SetAttributes("datasource.uid", firstNode.datasource.UID, attribute.Key("datasource.uid").String(firstNode.datasource.UID))
+			span.SetAttributes(
+				attribute.String("datasource.type", firstNode.datasource.Type),
+				attribute.String("datasource.uid", firstNode.datasource.UID),
+			)
 
 			req := &backend.QueryDataRequest{
 				PluginContext: pCtx,
@@ -261,11 +263,8 @@ func executeDSNodesGrouped(ctx context.Context, now time.Time, vars mathexp.Vars
 				if e != nil {
 					responseType = "error"
 					respStatus = "failure"
-					span.AddEvents([]string{"error", "message"},
-						[]tracing.EventValue{
-							{Str: fmt.Sprintf("%v", err)},
-							{Str: "failed to query data source"},
-						})
+					span.SetStatus(codes.Error, "failed to query data source")
+					span.RecordError(e)
 				}
 				logger.Debug("Data source queried", "responseType", responseType)
 				useDataplane := strings.HasPrefix(responseType, "dataplane-")
@@ -313,8 +312,10 @@ func (dn *DSNode) Execute(ctx context.Context, now time.Time, _ mathexp.Vars, s
 	if err != nil {
 		return mathexp.Results{}, err
 	}
-	span.SetAttributes("datasource.type", dn.datasource.Type, attribute.Key("datasource.type").String(dn.datasource.Type))
-	span.SetAttributes("datasource.uid", dn.datasource.UID, attribute.Key("datasource.uid").String(dn.datasource.UID))
+	span.SetAttributes(
+		attribute.String("datasource.type", dn.datasource.Type),
+		attribute.String("datasource.uid", dn.datasource.UID),
+	)
 
 	req := &backend.QueryDataRequest{
 		PluginContext: pCtx,
@@ -337,11 +338,8 @@ func (dn *DSNode) Execute(ctx context.Context, now time.Time, _ mathexp.Vars, s
 		if e != nil {
 			responseType = "error"
 			respStatus = "failure"
-			span.AddEvents([]string{"error", "message"},
-				[]tracing.EventValue{
-					{Str: fmt.Sprintf("%v", err)},
-					{Str: "failed to query data source"},
-				})
+			span.SetStatus(codes.Error, "failed to query data source")
+			span.RecordError(e)
 		}
 		logger.Debug("Data source queried", "responseType", responseType)
 		useDataplane := strings.HasPrefix(responseType, "dataplane-")

pkg/infra/httpclient/httpclientprovider/tracing_middleware.go

@@ -10,6 +10,7 @@ import (
 	"go.opentelemetry.io/contrib/instrumentation/net/http/httptrace/otelhttptrace"
 	"go.opentelemetry.io/otel/attribute"
 	"go.opentelemetry.io/otel/codes"
+	semconv "go.opentelemetry.io/otel/semconv/v1.17.0"
 	"go.opentelemetry.io/otel/trace"
 
 	"github.com/grafana/grafana/pkg/infra/log"
@@ -30,17 +31,18 @@ func TracingMiddleware(logger log.Logger, tracer tracing.Tracer) httpclient.Midd
 			ctx = httptrace.WithClientTrace(ctx, otelhttptrace.NewClientTrace(ctx, otelhttptrace.WithoutSubSpans(), otelhttptrace.WithoutHeaders()))
 			req = req.WithContext(ctx)
 			for k, v := range opts.Labels {
-				span.SetAttributes(k, v, attribute.Key(k).String(v))
+				span.SetAttributes(attribute.String(k, v))
 			}
 
 			tracer.Inject(ctx, req.Header, span)
 			res, err := next.RoundTrip(req)
 
-			span.SetAttributes("http.url", req.URL.String(), attribute.String("http.url", req.URL.String()))
-			span.SetAttributes("http.method", req.Method, attribute.String("http.method", req.Method))
+			span.SetAttributes(semconv.HTTPURL(req.URL.String()))
+			span.SetAttributes(semconv.HTTPMethod(req.Method))
 			// ext.SpanKind.Set(span, ext.SpanKindRPCClientEnum)
 
 			if err != nil {
+				span.SetStatus(codes.Error, "request failed")
 				span.RecordError(err)
 				return res, err
 			}
@@ -49,10 +51,10 @@ func TracingMiddleware(logger log.Logger, tracer tracing.Tracer) httpclient.Midd
 				// we avoid measuring contentlength less than zero because it indicates
 				// that the content size is unknown. https://godoc.org/github.com/badu/http#Response
 				if res.ContentLength > 0 {
-					span.SetAttributes(httpContentLengthTagKey, res.ContentLength, attribute.Key(httpContentLengthTagKey).Int64(res.ContentLength))
+					span.SetAttributes(attribute.Int64(httpContentLengthTagKey, res.ContentLength))
 				}
 
-				span.SetAttributes("http.status_code", res.StatusCode, attribute.Int("http.status_code", res.StatusCode))
+				span.SetAttributes(semconv.HTTPStatusCode(res.StatusCode))
 				if res.StatusCode >= 400 {
 					span.SetStatus(codes.Error, fmt.Sprintf("error with HTTP status code %s", strconv.Itoa(res.StatusCode)))
 				}

pkg/infra/serverlock/serverlock.go

@@ -34,7 +34,7 @@ type ServerLockService struct {
 func (sl *ServerLockService) LockAndExecute(ctx context.Context, actionName string, maxInterval time.Duration, fn func(ctx context.Context)) error {
 	start := time.Now()
 	ctx, span := sl.tracer.Start(ctx, "ServerLockService.LockAndExecute")
-	span.SetAttributes("serverlock.actionName", actionName, attribute.Key("serverlock.actionName").String(actionName))
+	span.SetAttributes(attribute.String("serverlock.actionName", actionName))
 	defer span.End()
 
 	ctxLogger := sl.log.FromContext(ctx)
@@ -138,7 +138,7 @@ func (sl *ServerLockService) getOrCreate(ctx context.Context, actionName string)
 func (sl *ServerLockService) LockExecuteAndRelease(ctx context.Context, actionName string, maxInterval time.Duration, fn func(ctx context.Context)) error {
 	start := time.Now()
 	ctx, span := sl.tracer.Start(ctx, "ServerLockService.LockExecuteAndRelease")
-	span.SetAttributes("serverlock.actionName", actionName, attribute.Key("serverlock.actionName").String(actionName))
+	span.SetAttributes(attribute.String("serverlock.actionName", actionName))
 	defer span.End()
 
 	ctxLogger := sl.log.FromContext(ctx)

pkg/infra/tracing/test_helper.go

@@ -24,7 +24,7 @@ func InitializeTracerForTest(opts ...TracerForTestOption) Tracer {
 
 	otel.SetTracerProvider(tp)
 
-	ots := &Opentelemetry{Propagation: "jaeger,w3c", tracerProvider: tp}
+	ots := &TracingService{Propagation: "jaeger,w3c", tracerProvider: tp}
 	_ = ots.initOpentelemetryTracer()
 	return ots
 }