Skip to content
Closed
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
Original file line number Diff line number Diff line change
Expand Up @@ -9,27 +9,26 @@
"strconv"
"strings"
"sync"
"sync/atomic"
"time"

backendconfig "github.com/rudderlabs/rudder-server/backend-config"
"github.com/rudderlabs/rudder-server/jsonrs"
transformerfs "github.com/rudderlabs/rudder-server/services/transformer"

obskit "github.com/rudderlabs/rudder-observability-kit/go/labels"

"github.com/cenkalti/backoff"
"github.com/samber/lo"

"github.com/rudderlabs/rudder-go-kit/config"
"github.com/rudderlabs/rudder-go-kit/filemanager"
"github.com/rudderlabs/rudder-go-kit/logger"
"github.com/rudderlabs/rudder-go-kit/stats"

obskit "github.com/rudderlabs/rudder-observability-kit/go/labels"
backendconfig "github.com/rudderlabs/rudder-server/backend-config"
transformerclient "github.com/rudderlabs/rudder-server/internal/transformer-client"
"github.com/rudderlabs/rudder-server/jsonrs"
"github.com/rudderlabs/rudder-server/processor/integrations"
transformerutils "github.com/rudderlabs/rudder-server/processor/internal/transformer"
"github.com/rudderlabs/rudder-server/processor/internal/transformer/destination_transformer/embedded/kafka"
"github.com/rudderlabs/rudder-server/processor/internal/transformer/destination_transformer/embedded/pubsub"
"github.com/rudderlabs/rudder-server/processor/types"
transformerfs "github.com/rudderlabs/rudder-server/services/transformer"
"github.com/rudderlabs/rudder-server/utils/httputil"
reportingtypes "github.com/rudderlabs/rudder-server/utils/types"
warehouseutils "github.com/rudderlabs/rudder-server/warehouse/utils"
Expand Down Expand Up @@ -78,9 +77,12 @@
handle.stats.matchedEvents = handle.stat.NewStat("embedded_destination_transform_matched_events", stats.CountType)
handle.stats.mismatchedEvents = handle.stat.NewStat("embedded_destination_transform_mismatched_events", stats.CountType)

handle.loggedEvents = 0
handle.loggedEventsMu = sync.Mutex{}
handle.loggedFileName = generateLogFileName()
var err error
handle.samplingFileManager, err = getSamplingUploader(conf, log)
if err != nil {
log.Errorn("failed to create dt sampling file manager", obskit.Error(err))
handle.samplingFileManager = nil
}

Check warning on line 85 in processor/internal/transformer/destination_transformer/destination_transformer.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/destination_transformer.go#L83-L85

Added lines #L83 - L85 were not covered by tests

handle.config.compactionEnabled = conf.GetReloadableBoolVar(false, "Processor.DestinationTransformer.compactionEnabled", "Transformer.compactionEnabled")

Expand Down Expand Up @@ -116,9 +118,8 @@
mismatchedEvents stats.Counter
}

loggedEventsMu sync.Mutex
loggedEvents int64
loggedFileName string
loggedEvents atomic.Int64
samplingFileManager *filemanager.S3Manager
}

func (d *Client) transform(ctx context.Context, clientEvents []types.TransformerEvent) types.Response {
Expand Down Expand Up @@ -391,7 +392,9 @@
legacyTransformerResponse := c.transform(ctx, clientEvents)
embeddedTransformerResponse := impl(ctx, clientEvents)

c.CompareAndLog(embeddedTransformerResponse, legacyTransformerResponse)
go func() {
c.CompareAndLog(ctx, embeddedTransformerResponse, legacyTransformerResponse)
}()

Check warning on line 397 in processor/internal/transformer/destination_transformer/destination_transformer.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/destination_transformer.go#L395-L397

Added lines #L395 - L397 were not covered by tests

return legacyTransformerResponse
}
Expand Down Expand Up @@ -429,3 +432,32 @@
}
return jsonrs.Marshal(data)
}

func getSamplingUploader(conf *config.Config, log logger.Logger) (*filemanager.S3Manager, error) {
var (
bucket = conf.GetString("DTSampling.Bucket", "processor-dt-sampling")
endpoint = conf.GetString("DTSampling.Endpoint", "")
accessKeyID = conf.GetStringVar("", "DTSampling.AccessKeyId", "AWS_ACCESS_KEY_ID")
accessKey = conf.GetStringVar("", "DTSampling.AccessKey", "AWS_SECRET_ACCESS_KEY")
s3ForcePathStyle = conf.GetBool("DTSampling.S3ForcePathStyle", false)
disableSSL = conf.GetBool("DTSampling.DisableSsl", false)
enableSSE = conf.GetBoolVar(false, "DTSampling.EnableSse", "AWS_ENABLE_SSE")
useGlue = conf.GetBool("DTSampling.UseGlue", false)
region = conf.GetStringVar("us-east-1", "DTSampling.Region", "AWS_DEFAULT_REGION")
)
s3Config := map[string]any{
"bucketName": bucket,
"endpoint": endpoint,
"accessKeyID": accessKeyID,
"accessKey": accessKey,
"s3ForcePathStyle": s3ForcePathStyle,
"disableSSL": disableSSL,
"enableSSE": enableSSE,
"useGlue": useGlue,
"region": region,
}

return filemanager.NewS3Manager(s3Config, log.Withn(logger.NewStringField("component", "dt-uploader")), func() time.Duration {
return conf.GetDuration("DTSampling.Timeout", 120, time.Second)
})

Check warning on line 462 in processor/internal/transformer/destination_transformer/destination_transformer.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/destination_transformer.go#L461-L462

Added lines #L461 - L462 were not covered by tests
}
63 changes: 29 additions & 34 deletions processor/internal/transformer/destination_transformer/logger.go
Original file line number Diff line number Diff line change
@@ -1,48 +1,62 @@
package destination_transformer

import (
"bytes"
"context"
"fmt"
"path"

"github.com/google/go-cmp/cmp"
"github.com/google/uuid"
"github.com/samber/lo"

"github.com/rudderlabs/rudder-go-kit/config"
"github.com/rudderlabs/rudder-go-kit/logger"
obskit "github.com/rudderlabs/rudder-observability-kit/go/labels"

"github.com/rudderlabs/rudder-go-kit/stringify"

"github.com/rudderlabs/rudder-server/jsonrs"
"github.com/rudderlabs/rudder-server/processor/types"
"github.com/rudderlabs/rudder-server/utils/misc"
)

func (c *Client) CompareAndLog(
ctx context.Context,
embeddedResponse, legacyResponse types.Response,
) {
c.loggedEventsMu.Lock()
defer c.loggedEventsMu.Unlock()
if c.samplingFileManager == nil { // Cannot upload, we should just report the issue with no diff
c.log.Errorn("DestinationTransformer sanity check failed")
return
}

Check warning on line 27 in processor/internal/transformer/destination_transformer/logger.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/logger.go#L24-L27

Added lines #L24 - L27 were not covered by tests

if c.loggedEvents >= int64(c.config.maxLoggedEvents.Load()) {
if c.loggedEvents.Load() >= int64(c.config.maxLoggedEvents.Load()) {

Check warning on line 29 in processor/internal/transformer/destination_transformer/logger.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/logger.go#L29

Added line #L29 was not covered by tests
return
}

c.stats.comparisonTime.RecordDuration()()

differingResponse, sampleDiff := c.differingEvents(embeddedResponse, legacyResponse)
if len(differingResponse) == 0 && sampleDiff == "" {
c.log.Infof("Embedded and legacy responses are matches")
return
}

logEntries := lo.Map(differingResponse, func(item types.TransformerResponse, index int) string {
return stringify.Any(item)
})
if err := c.write(append([]string{sampleDiff}, logEntries...)); err != nil {
c.log.Warnn("Error logging events", obskit.Error(err))
objName := path.Join("embedded-dt-samples", config.GetKubeNamespace(), uuid.New().String())
differingResponseJSON, err := jsonrs.Marshal(differingResponse)
if err != nil {
c.log.Errorn("DestinationTransformer sanity check failed (cannot encode differingResponse)", obskit.Error(err))
return
}

Check warning on line 45 in processor/internal/transformer/destination_transformer/logger.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/logger.go#L40-L45

Added lines #L40 - L45 were not covered by tests

// upload sample diff and differing response to s3
file, err := c.samplingFileManager.UploadReader(ctx, objName, bytes.NewReader(append([]byte(sampleDiff), differingResponseJSON...)))
if err != nil {
c.log.Errorn("Error uploading DestinationTransformer sanity check diff file", obskit.Error(err))

Check warning on line 50 in processor/internal/transformer/destination_transformer/logger.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/logger.go#L48-L50

Added lines #L48 - L50 were not covered by tests
return
}

c.log.Infof("Successfully logged events: %d", len(logEntries))
c.loggedEvents += int64(len(logEntries))
c.log.Errorn("DestinationTransformer sanity check failed",
logger.NewStringField("location", file.Location),
logger.NewStringField("objectName", file.ObjectName),
)

c.loggedEvents.And(int64(len(differingResponse)))

Check warning on line 59 in processor/internal/transformer/destination_transformer/logger.go

View check run for this annotation

Codecov / codecov/patch

processor/internal/transformer/destination_transformer/logger.go#L54-L59

Added lines #L54 - L59 were not covered by tests
}

func (c *Client) differingEvents(
Expand Down Expand Up @@ -95,22 +109,3 @@
c.stats.mismatchedEvents.Count(differedEventsCount)
return differedSampleEvents, sampleDiff
}

func (c *Client) write(data []string) error {
writer, err := misc.CreateGZ(c.loggedFileName)
if err != nil {
return fmt.Errorf("creating buffered writer: %w", err)
}
defer func() { _ = writer.Close() }()

for _, entry := range data {
if _, err := writer.Write([]byte(entry + "\n")); err != nil {
return fmt.Errorf("writing log entry: %w", err)
}
}
return nil
}

func generateLogFileName() string {
return fmt.Sprintf("destination_transformations_debug_%s.log.gz", uuid.NewString())
}
Loading