Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions internal/tasktelemetry/otlp/exporter.go
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@ import (
"net/url"
"time"

"github.com/rs/zerolog/log"
"github.com/warpdotdev/oz-agent-worker/internal/types"
"go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploghttp"
"go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp"
Expand Down Expand Up @@ -108,6 +109,8 @@ type logExporter struct {
func (e logExporter) Export(ctx context.Context, records []sdklog.Record) error {
// OTel reports exporter errors outside the task-local log redactor.
if err := e.Exporter.Export(ctx, records); err != nil {
// The context logger would send the diagnostic back to the task collector.
log.Error().Err(err).Msg("task log export failed")
return errors.New("task log export failed")
}
return nil
Expand Down
60 changes: 60 additions & 0 deletions internal/tasktelemetry/otlp/exporter_test.go
Original file line number Diff line number Diff line change
@@ -1,16 +1,21 @@
package otlp

import (
"bytes"
"context"
"encoding/json"
"encoding/pem"
"net/http"
"net/http/httptest"
"os"
"path/filepath"
"strings"
"sync/atomic"
"testing"
"time"

"github.com/rs/zerolog"
"github.com/rs/zerolog/log"
"github.com/warpdotdev/oz-agent-worker/internal/types"
"go.opentelemetry.io/otel/attribute"
sdklog "go.opentelemetry.io/otel/sdk/log"
Expand All @@ -20,6 +25,61 @@ import (
"golang.org/x/oauth2"
)

func TestLogExportErrorStaysInWorkerLog(t *testing.T) {
const diagnostic = "collector rejected the export: diagnostic detail"
var requests atomic.Int32
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, req *http.Request) {
requests.Add(1)
http.Error(w, diagnostic, http.StatusForbidden)
}))
defer server.Close()

var workerOutput, taskOutput bytes.Buffer
previous := log.Logger
log.Logger = zerolog.New(&workerOutput)
t.Cleanup(func() { log.Logger = previous })
ctx := zerolog.New(&taskOutput).WithContext(t.Context())
config, err := NewExportConfig(ctx, "", &types.TaskAssignmentMessage{
TelemetryCollection: &types.TelemetryCollectionConfig{
Endpoint: server.URL,
BootstrapToken: &types.TelemetryBootstrapToken{
Token: "task-token", ExpiresAt: time.Now().Add(time.Hour).Format(time.RFC3339Nano),
},
},
})
if err != nil {
t.Fatal(err)
}
exporter, err := config.LogExporter()
if err != nil {
t.Fatal(err)
}
defer func() { _ = exporter.Shutdown(context.Background()) }()
var record sdklog.Record
record.SetBody(attribute.StringValue("task output"))
err = exporter.Export(ctx, []sdklog.Record{record})
if err == nil || err.Error() != "task log export failed" {
t.Fatalf("exporter returned an unsafe error: %v", err)
}
if taskOutput.Len() != 0 {
t.Fatalf("exporter diagnostic reached the task logger: %s", &taskOutput)
}
var event struct {
Level string `json:"level"`
Error string `json:"error"`
Message string `json:"message"`
}
if err := json.Unmarshal(workerOutput.Bytes(), &event); err != nil {
t.Fatalf("missing worker diagnostic: %v", err)
}
if event.Level != "error" || !strings.Contains(event.Error, diagnostic) || event.Message != "task log export failed" {
t.Fatalf("worker diagnostic lost the original error: %+v", event)
}
if requests.Load() != 1 {
t.Fatalf("exporter made %d requests, want one failed export", requests.Load())
}
}

func TestHTTPSExport(t *testing.T) {
for _, signal := range []string{"traces", "logs"} {
t.Run(signal, func(t *testing.T) {
Expand Down
Loading