Warning: The OpenCensus project is obsolete and was archived on July 31st, 2023. This means that any security vulnerabilities that are found will not be patched. We recommend that you migrate from OpenCensus tracing to OpenTelemetry, the successor project. See OpenCensus below for details.
Logging, debugging and telemetry all capture data that can be used for troubleshooting. Logging records specific events and transactions. Debugging exposes values for immediate analysis. Telemetry is suitable for production use and can serve both logging and monitoring purposes. Telemetry tracing follows requests through a system to provide a view of component interactions. Telemetry metrics collects data for significant performance indicators, offering insights into a system's health.
While working with the Go Client Libraries you may run into some situations where you need a deeper level of understanding about what is going on in order to solve your problem. Here are some tips and tricks that you can use in these cases. Note that many of the tips in this section will have a performance impact and are therefore not recommended for sustained production use. Use these tips locally or in production for a limited time to help get a better understanding of what is going on.
All of our auto-generated clients have a constructor to create a client that uses HTTP/JSON instead of gRPC. Additionally a couple of our hand-written clients like Storage and Bigquery are also HTTP based. Here are some tips for debugging these clients.
Try setting the following environment variable for verbose Go HTTP logging: GODEBUG=http2debug=1. To read more about this feature please see the godoc for net/http.
WARNING: Enabling this debug variable will log headers and payloads which may contain private information.
You may want to add in your own logging around HTTP requests. One way to do this is to register a custom HTTP client with a logging transport built in. Here is an example of how you would do this with the storage client.
WARNING: Adding this middleware will log headers and payloads which may contain private information.
package main
import (
"context"
"fmt"
"log"
"net/http"
"net/http/httputil"
"cloud.google.com/go/storage"
"google.golang.org/api/iterator"
"google.golang.org/api/option"
htransport "google.golang.org/api/transport/http"
)
type loggingRoundTripper struct {
rt http.RoundTripper
}
func (d loggingRoundTripper) RoundTrip(r *http.Request) (*http.Response, error) {
// Will create a dump of the request and body.
dump, err := httputil.DumpRequest(r, true)
if err != nil {
log.Println("error dumping request")
}
log.Printf("%s", dump)
return d.rt.RoundTrip(r)
}
func main() {
ctx := context.Background()
// Create a transport with authentication built-in detected with
// [ADC](https://google.aip.dev/auth/4110). Note you will have to pass any
// required scoped for the client you are using.
trans, err := htransport.NewTransport(ctx,
http.DefaultTransport,
option.WithScopes(storage.ScopeFullControl),
)
if err != nil {
log.Fatal(err)
}
// Embed customized transport into an HTTP client.
hc := &http.Client{
Transport: loggingRoundTripper{rt: trans},
}
// Supply custom HTTP client for use by the library.
client, err := storage.NewClient(ctx, option.WithHTTPClient(hc))
if err != nil {
log.Fatal(err)
}
defer client.Close()
// Use the client
}Try setting the following environment variables for grpc-go:
GRPC_GO_LOG_VERBOSITY_LEVEL=99 GRPC_GO_LOG_SEVERITY_LEVEL=info. These are
good for diagnosing connection level failures. For more information please see
grpc-go's debug documentation.
You may want to add in your own logging around gRPC requests. One way to do this is to register a custom interceptor that adds logging. Here is an example of how you would do this with the secretmanager client. Note this example registers a UnaryClientInterceptor but you may want/need to register a StreamClientInterceptor instead-of/as-well depending on what kinds of RPCs you are calling.
WARNING: Adding this interceptor will log metadata and payloads which may contain private information.
package main
import (
"context"
"log"
secretmanager "cloud.google.com/go/secretmanager/apiv1"
"google.golang.org/api/option"
"google.golang.org/grpc"
"google.golang.org/grpc/metadata"
"google.golang.org/protobuf/encoding/protojson"
"google.golang.org/protobuf/reflect/protoreflect"
)
func loggingUnaryInterceptor() grpc.UnaryClientInterceptor {
return func(ctx context.Context, method string, req, reply interface{}, cc *grpc.ClientConn, invoker grpc.UnaryInvoker, opts ...grpc.CallOption) error {
err := invoker(ctx, method, req, reply, cc, opts...)
log.Printf("Invoked method: %v", method)
md, ok := metadata.FromOutgoingContext(ctx)
if ok {
log.Println("Metadata:")
for k, v := range md {
log.Printf("Key: %v, Value: %v", k, v)
}
}
reqb, merr := protojson.Marshal(req.(protoreflect.ProtoMessage))
if merr == nil {
log.Printf("Request: %s", reqb)
}
return err
}
}
func main() {
ctx := context.Background()
// Supply custom gRPC interceptor for use by the client.
client, err := secretmanager.NewClient(ctx,
option.WithGRPCDialOption(grpc.WithUnaryInterceptor(loggingUnaryInterceptor())),
)
if err != nil {
log.Fatal(err)
}
defer client.Close()
// Use the client
}Warning: The OpenCensus project is obsolete and was archived on July 31st, 2023. This means that any security vulnerabilities that are found will not be patched. We recommend that you migrate from OpenCensus tracing to OpenTelemetry, the successor project. The default experimental tracing support for OpenCensus is now deprecated in the Google Cloud client libraries for Go. See OpenCensus below for details.
The Google Cloud client libraries for Go now use the OpenTelemetry project by default. Temporary opt-in support for OpenCensus is still available. The transition from OpenCensus to OpenTelemetry is covered in the following sections.
Apart from spans created by underlying libraries such as gRPC, Google Cloud Go generated clients do not create spans. Only the spans created by following hand-written clients are in scope for the discussion in this section:
- cloud.google.com/go/bigquery
- cloud.google.com/go/bigtable
- cloud.google.com/go/datastore
- cloud.google.com/go/firestore
- cloud.google.com/go/spanner
- cloud.google.com/go/storage
Currently, the spans created by these clients are for OpenTelemetry. OpenCensus users are urged to transition to OpenTelemetry as soon as possible, as explained in the next section. OpenCensus users can still opt-in to the deprecated OpenCensus support via an environment variable, as described below.
Warning: The OpenCensus project is obsolete and was archived on July 31st, 2023. This means that any security vulnerabilities that are found will not be patched. We recommend that you migrate from OpenCensus tracing to OpenTelemetry, the successor project. The default experimental tracing support for OpenCensus is now deprecated in the Google Cloud client libraries for Go.
Using the OpenTelemetry-Go - OpenCensus Bridge, you can immediately begin exporting your traces with OpenTelemetry, even while dependencies of your application remain instrumented with OpenCensus. If you do not use the bridge, you will need to migrate your entire application and all of its instrumented dependencies at once. For simple applications, this may be possible, but we expect the bridge to be helpful if multiple libraries with instrumentation are used.
On May 29, 2024, six months after the
release
of experimental, opt-in support for OpenTelemetry tracing, the default tracing
support in the clients above was changed from OpenCensus to OpenTelemetry, and
the experimental OpenCensus support was marked as deprecated. To continue
using the OpenCensus support, set the environment variable
GOOGLE_API_GO_EXPERIMENTAL_TELEMETRY_PLATFORM_TRACING to the case-insensitive
value opencensus before loading the client library.
export GOOGLE_API_GO_EXPERIMENTAL_TELEMETRY_PLATFORM_TRACING=opencensusOn December 2nd, 2024, one year after the release of OpenTelemetry support, the experimental and deprecated support for OpenCensus tracing will be removed.
Please note that all Google Cloud Go clients currently provide experimental support for the propagation of both OpenCensus and OpenTelemetry trace context to their receiving endpoints. The experimental support for OpenCensus trace context propagation will be removed at the same time as the experimental OpenCensus tracing support.
Please refer to the following resources:
The default experimental tracing support for OpenCensus is now deprecated in the Google Cloud client libraries for Go.
On May 29, 2024, the default experimental tracing support in the Google Cloud client libraries for Go was changed from OpenCensus to OpenTelemetry.
Warning: OpenTelemetry-Go ensures compatibility with ONLY the current supported versions of the Go language. This support may be narrower than the support that has been offered historically by the Go Client Libraries. Ensure that your Go runtime version is supported by the OpenTelemetry-Go compatibility policy before enabling OpenTelemetry instrumentation.
Please refer to the following resources:
To configure the OpenCensus bridge with OpenTelemetry and Cloud Trace:
import (
"context"
"log"
"os"
texporter "github.com/GoogleCloudPlatform/opentelemetry-operations-go/exporter/trace"
octrace "go.opencensus.io/trace"
"go.opentelemetry.io/contrib/detectors/gcp"
"go.opentelemetry.io/otel"
"go.opentelemetry.io/otel/bridge/opencensus"
"go.opentelemetry.io/otel/sdk/resource"
sdktrace "go.opentelemetry.io/otel/sdk/trace"
semconv "go.opentelemetry.io/otel/semconv/v1.7.0"
)
func main() {
// Create exporter.
ctx := context.Background()
projectID := os.Getenv("GOOGLE_CLOUD_PROJECT")
exporter, err := texporter.New(texporter.WithProjectID(projectID))
if err != nil {
log.Fatalf("texporter.New: %v", err)
}
// Identify your application using resource detection
res, err := resource.New(ctx,
// Use the GCP resource detector to detect information about the GCP platform
resource.WithDetectors(gcp.NewDetector()),
// Keep the default detectors
resource.WithTelemetrySDK(),
// Add your own custom attributes to identify your application
resource.WithAttributes(
semconv.ServiceNameKey.String("my-application"),
),
)
if err != nil {
log.Fatalf("resource.New: %v", err)
}
// Create trace provider with the exporter.
//
// By default it uses AlwaysSample() which samples all traces.
// In a production environment or high QPS setup please use
// probabilistic sampling.
// Example:
// tp := sdktrace.NewTracerProvider(sdktrace.WithSampler(sdktrace.TraceIDRatioBased(0.0001)), ...)
tp := sdktrace.NewTracerProvider(
sdktrace.WithBatcher(exporter),
sdktrace.WithResource(res),
)
defer tp.Shutdown(ctx) // flushes any pending spans, and closes connections.
otel.SetTracerProvider(tp)
tracer := otel.GetTracerProvider().Tracer("example.com/trace")
// Configure the OpenCensus tracer to use the bridge.
octrace.DefaultTracer = opencensus.NewTracer(tracer)
// Use otel tracer to create spans...
}In order to pass options to OpenTelemetry trace context propagation, follow the appropriate example for the client's underlying transport.
ctx := context.Background()
trans, err := htransport.NewTransport(ctx,
http.DefaultTransport,
option.WithScopes(storage.ScopeFullControl),
)
if err != nil {
log.Fatal(err)
}
// An example of passing options to the otelhttp.Transport.
otelOpts := otelhttp.WithFilter(func(r *http.Request) bool {
return r.URL.Path != "/ping"
})
hc := &http.Client{
Transport: otelhttp.NewTransport(trans, otelOpts),
}
client, err := storage.NewClient(ctx, option.WithHTTPClient(hc))Note that scopes must be set manually in this user-configured solution.
projectID := "..."
ctx := context.Background()
// An example of passing options to grpc.WithStatsHandler.
otelOpts := otelgrpc.WithMessageEvents(otelgrpc.ReceivedEvents)
dialOpts := grpc.WithStatsHandler(otelgrpc.NewClientHandler(otelOpts))
ctx := context.Background()
c, err := datastore.NewClient(ctx, projectID, option.WithGRPCDialOption(dialOpts))
if err != nil {
log.Fatal(err)
}
defer c.Close()The generated clients do not create metrics. Only the following hand-written clients create experimental OpenCensus metrics:
The transition of the experimental metrics in the clients above from OpenCensus to OpenTelemetry is still TBD.