Compare commits

...

1 Commits

Author SHA1 Message Date
Tushar Vats
be7df2f8f7 refactor(logparsingpipeline): remove normalize operator
signozlogsnormalizer in the collector normalizes log bodies, so the server
no longer injects the stanza normalize operator under use_json_body.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-13 21:39:02 +05:30
8 changed files with 9 additions and 192 deletions

View File

@@ -56,7 +56,6 @@ func NewServer(config signoz.Config, signoz *signoz.SigNoz) (*Server, error) {
signoz.SQLStore,
integrationsController.GetPipelinesForInstalledIntegrations,
reader,
signoz.Flagger,
)
if err != nil {
return nil, err

View File

@@ -2688,19 +2688,7 @@ func (aH *APIHandler) PreviewLogsPipelinesHandler(w http.ResponseWriter, r *http
return
}
claims, errv2 := authtypes.ClaimsFromContext(r.Context())
if errv2 != nil {
render.Error(w, errv2)
return
}
orgID, errv2 := valuer.NewUUID(claims.OrgID)
if errv2 != nil {
render.Error(w, errv2)
return
}
resultLogs, err := aH.LogsParsingPipelineController.PreviewLogsPipelines(r.Context(), orgID, &req)
resultLogs, err := aH.LogsParsingPipelineController.PreviewLogsPipelines(r.Context(), &req)
if err != nil {
render.Error(w, err)
return

View File

@@ -165,22 +165,6 @@ func isNormalizerProcessor(name string) bool {
return name == normalizerProcessor || strings.HasPrefix(name, normalizerProcessorPrefix)
}
// collectorNormalizesFirst reports whether the logs pipeline of the collector config runs
// signozlogsnormalizer ahead of user pipelines, judged by the first one as in buildCollectorPipelineProcessorsList.
// Malformed configs report false; generateCollectorConfigWithPipelines surfaces their errors.
func collectorNormalizesFirst(collectorConf map[string]interface{}) bool {
p, err := getOtelPipelineFromConfig(collectorConf)
if err != nil || p.Pipelines.Logs == nil {
return false
}
normalizerIdx := slices.IndexFunc(p.Pipelines.Logs.Processors, isNormalizerProcessor)
if normalizerIdx < 0 {
return false
}
processorConfigs, _ := collectorConf["processors"].(map[string]interface{})
return opampPosition(processorConfigs, p.Pipelines.Logs.Processors[normalizerIdx]) == normalizerBeforeUserPipelines
}
func opampPosition(processorConfigs map[string]interface{}, name string) string {
config, _ := processorConfigs[name].(map[string]interface{})
opamp, _ := config["opamp"].(map[string]interface{})

View File

@@ -5,12 +5,10 @@ import (
"fmt"
"testing"
"github.com/SigNoz/signoz/pkg/flagger/flaggertest"
"github.com/SigNoz/signoz/pkg/query-service/constants"
"github.com/SigNoz/signoz/pkg/query-service/model"
v3 "github.com/SigNoz/signoz/pkg/query-service/model/v3"
"github.com/SigNoz/signoz/pkg/types/pipelinetypes"
"github.com/SigNoz/signoz/pkg/valuer"
. "github.com/smartystreets/goconvey/convey"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
@@ -501,61 +499,3 @@ func TestPipeCharInAliasDoesntBreakCollectorConfig(t *testing.T) {
}, result[0].Attributes_string,
)
}
func TestRecommendAgentConfigWithNormalizerProcessorAndJSONBody(t *testing.T) {
userPipeline := makeTestAddAttributePipeline()
controller := &LogParsingPipelineController{
fl: flaggertest.WithUseJSONBody(t, true),
GetIntegrationPipelines: func(context.Context, string) ([]pipelinetypes.GettablePipeline, error) {
return []pipelinetypes.GettablePipeline{userPipeline}, nil
},
}
normalizeOperator := CollectorConfProcessorName(getNormalizePipeline())
user := CollectorConfProcessorName(userPipeline)
testCases := []struct {
name string
processors string
logs string
expected []any
}{
{
name: "NoNormalizer_OperatorPrepended",
logs: "[memory_limiter, batch]",
expected: []any{"memory_limiter", normalizeOperator, user, "batch"},
},
{
name: "NormalizerDefault_OperatorKept",
logs: "[memory_limiter, signozlogsnormalizer, batch]",
expected: []any{"memory_limiter", normalizeOperator, user, "signozlogsnormalizer", "batch"},
},
{
name: "NormalizerBefore_OperatorDropped",
processors: "signozlogsnormalizer:\n opamp:\n position: before\n",
logs: "[memory_limiter, signozlogsnormalizer, batch]",
expected: []any{"memory_limiter", "signozlogsnormalizer", user, "batch"},
},
{
name: "NormalizerAfter_OperatorKept",
processors: "signozlogsnormalizer:\n opamp:\n position: after\n",
logs: "[memory_limiter, signozlogsnormalizer, batch]",
expected: []any{"memory_limiter", normalizeOperator, user, "signozlogsnormalizer", "batch"},
},
}
for _, testCase := range testCases {
t.Run(testCase.name, func(t *testing.T) {
conf := fmt.Sprintf(
"processors:\n memory_limiter: {}\n batch: {}\n %s\nservice:\n pipelines:\n logs:\n receivers: [otlp]\n processors: %s\n exporters: [clickhouselogsexporter]\n",
testCase.processors, testCase.logs,
)
recommended, _, err := controller.RecommendAgentConfig(valuer.GenerateUUID(), []byte(conf), nil)
require.NoError(t, err)
var recommendedConf map[string]any
require.NoError(t, yaml.Unmarshal(recommended, &recommendedConf))
logs := recommendedConf["service"].(map[string]any)["pipelines"].(map[string]any)["logs"].(map[string]any)
assert.Equal(t, testCase.expected, logs["processors"])
})
}
}

View File

@@ -11,7 +11,6 @@ import (
"github.com/google/uuid"
"github.com/SigNoz/signoz/pkg/errors"
"github.com/SigNoz/signoz/pkg/flagger"
"github.com/SigNoz/signoz/pkg/query-service/agentConf"
"github.com/SigNoz/signoz/pkg/query-service/constants"
"github.com/SigNoz/signoz/pkg/query-service/interfaces"
@@ -20,7 +19,6 @@ import (
"github.com/SigNoz/signoz/pkg/query-service/utils"
"github.com/SigNoz/signoz/pkg/sqlstore"
"github.com/SigNoz/signoz/pkg/types"
"github.com/SigNoz/signoz/pkg/types/featuretypes"
"github.com/SigNoz/signoz/pkg/types/opamptypes"
"github.com/SigNoz/signoz/pkg/types/pipelinetypes"
"github.com/SigNoz/signoz/pkg/valuer"
@@ -39,21 +37,18 @@ type LogParsingPipelineController struct {
GetIntegrationPipelines func(context.Context, string) ([]pipelinetypes.GettablePipeline, error)
// TODO(Piyush): remove with qbv5 migration
reader interfaces.Reader
fl flagger.Flagger
}
func NewLogParsingPipelinesController(
sqlStore sqlstore.SQLStore,
getIntegrationPipelines func(context.Context, string) ([]pipelinetypes.GettablePipeline, error),
reader interfaces.Reader,
fl flagger.Flagger,
) (*LogParsingPipelineController, error) {
repo := NewRepo(sqlStore)
return &LogParsingPipelineController{
Repo: repo,
GetIntegrationPipelines: getIntegrationPipelines,
reader: reader,
fl: fl,
}, nil
}
@@ -220,44 +215,6 @@ func (ic *LogParsingPipelineController) ValidatePipelines(ctx context.Context,
return err
}
// withNormalizePipeline places normalize ahead of user pipelines when queries run on body_v2
// (use_json_body), so operators see the body the explorer shows. Absent when the flag is off
// or collectorNormalizesFirst.
func (ic *LogParsingPipelineController) withNormalizePipeline(ctx context.Context, orgID valuer.UUID, pipelines []pipelinetypes.GettablePipeline, collectorNormalizesFirst bool) []pipelinetypes.GettablePipeline {
if collectorNormalizesFirst || !ic.fl.BooleanOrEmpty(ctx, flagger.FeatureUseJSONBody, featuretypes.NewFlaggerEvaluationContext(orgID)) {
return pipelines
}
return append([]pipelinetypes.GettablePipeline{getNormalizePipeline()}, pipelines...)
}
func getNormalizePipeline() pipelinetypes.GettablePipeline {
return pipelinetypes.GettablePipeline{
StoreablePipeline: pipelinetypes.StoreablePipeline{
Name: "Default Pipeline - PreProcessing Body",
Alias: "NormalizeBodyDefault",
Enabled: true,
},
Filter: &v3.FilterSet{
Items: []v3.FilterItem{
{
Key: v3.AttributeKey{
Key: "body",
},
Operator: v3.FilterOperatorExists,
},
},
},
Config: []pipelinetypes.PipelineOperator{
{
ID: uuid.NewString(),
Type: "normalize",
Enabled: true,
If: "body != nil",
},
},
}
}
// Returns effective list of pipelines including user created
// pipelines and pipelines for installed integrations
func (ic *LogParsingPipelineController) getEffectivePipelinesByVersion(
@@ -352,7 +309,6 @@ type PipelinesPreviewResponse struct {
func (ic *LogParsingPipelineController) PreviewLogsPipelines(
ctx context.Context,
orgID valuer.UUID,
request *PipelinesPreviewRequest,
) (*PipelinesPreviewResponse, error) {
pipelines, err := ic.enrichPipelinesFilters(ctx, request.Pipelines)
@@ -360,11 +316,6 @@ func (ic *LogParsingPipelineController) PreviewLogsPipelines(
return nil, err
}
// The collector gets the same pipeline prepended over opamp; see RecommendAgentConfig.
if ic.fl.BooleanOrEmpty(ctx, flagger.FeatureUseJSONBody, featuretypes.NewFlaggerEvaluationContext(orgID)) {
pipelines = append([]pipelinetypes.GettablePipeline{getNormalizePipeline()}, pipelines...)
}
result, collectorLogs, err := SimulatePipelinesProcessing(ctx, pipelines, request.Logs)
if err != nil {
return nil, err
@@ -383,16 +334,13 @@ func (pc *LogParsingPipelineController) AgentFeatureType() agentConf.AgentFeatur
// Implements agentConf.AgentFeature interface.
// RecommendAgentConfig generates the collector config to be sent to agents.
// The normalize pipeline (when use_json_body is on) is placed
// here, after rawPipelineData is serialized. So it is only present in the config sent to
// the collector and never persisted to the database as part of the user's pipeline list.
//
// NOTE: The configId sent to agents is derived from the pipeline version number
// (e.g. "LogPipelines:5"), not the YAML content. If server-side logic changes
// the generated YAML without bumping the version (e.g. toggling the use_json_body
// flag or updating operator IfExpressions), agents that already
// applied that version will not re-apply the new config. In such cases, users must save a
// new pipeline version via the API to force agents to pick up the change.
// the generated YAML without bumping the version (e.g. updating operator IfExpressions),
// agents that already applied that version will not re-apply the new config. In such
// cases, users must save a new pipeline version via the API to force agents to pick up
// the change.
func (pc *LogParsingPipelineController) RecommendAgentConfig(
orgId valuer.UUID,
currentConfYaml []byte,
@@ -418,15 +366,7 @@ func (pc *LogParsingPipelineController) RecommendAgentConfig(
return nil, "", err
}
collectorConf, err := unmarshalCollectorConfig(currentConfYaml)
if err != nil {
return nil, "", err
}
// normalize is only for sending to the collector, never persisted
enrichedPipelines = pc.withNormalizePipeline(ctx, orgId, enrichedPipelines, collectorNormalizesFirst(collectorConf))
updatedConf, err := generateCollectorConfigWithPipelines(collectorConf, enrichedPipelines)
updatedConf, err := GenerateCollectorConfigWithPipelines(currentConfYaml, enrichedPipelines)
if err != nil {
return nil, "", err
}

View File

@@ -6,11 +6,9 @@ import (
"testing"
"time"
"github.com/SigNoz/signoz/pkg/flagger/flaggertest"
"github.com/SigNoz/signoz/pkg/query-service/model"
v3 "github.com/SigNoz/signoz/pkg/query-service/model/v3"
"github.com/SigNoz/signoz/pkg/types/pipelinetypes"
"github.com/SigNoz/signoz/pkg/valuer"
"github.com/google/uuid"
"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/stanza/entry"
"github.com/stretchr/testify/assert"
@@ -119,42 +117,11 @@ func TestPipelinePreview(t *testing.T) {
}
func TestPipelinePreviewNormalizesBodyWithJSONBodyEnabled(t *testing.T) {
controller := &LogParsingPipelineController{fl: flaggertest.WithUseJSONBody(t, true)}
func TestPipelinePreviewKeepsBodyAsIs(t *testing.T) {
controller := &LogParsingPipelineController{}
result, err := controller.PreviewLogsPipelines(
context.Background(),
valuer.GenerateUUID(),
&PipelinesPreviewRequest{
Pipelines: []pipelinetypes.GettablePipeline{makeTestAddAttributePipeline()},
Logs: []model.SignozLog{
makeTestSignozLog("test log body", map[string]interface{}{"method": "GET"}),
makeTestSignozLog(
`{"level":"error","msg":"json log body"}`,
map[string]interface{}{"method": "GET"},
),
},
},
)
require.NoError(t, err)
require.Len(t, result.OutputLogs, 2)
assert.Equal(t, `{"message":"test log body"}`, result.OutputLogs[0].Body)
assert.Equal(
t,
`{"level":"error","message":"json log body"}`,
result.OutputLogs[1].Body,
)
assert.Equal(t, "val", result.OutputLogs[0].Attributes_string["test"])
}
func TestPipelinePreviewKeepsBodyAsIsWithJSONBodyDisabled(t *testing.T) {
controller := &LogParsingPipelineController{fl: flaggertest.WithUseJSONBody(t, false)}
result, err := controller.PreviewLogsPipelines(
context.Background(),
valuer.GenerateUUID(),
&PipelinesPreviewRequest{
Pipelines: []pipelinetypes.GettablePipeline{makeTestAddAttributePipeline()},
Logs: []model.SignozLog{

View File

@@ -46,7 +46,6 @@ func NewServer(config signoz.Config, signoz *signoz.SigNoz) (*Server, error) {
signoz.SQLStore,
integrationsController.GetPipelinesForInstalledIntegrations,
reader,
signoz.Flagger,
)
if err != nil {
return nil, err

View File

@@ -174,7 +174,7 @@ class Logs(ABC):
self.body_v2 = body_v2
else:
# Try to parse body as JSON; if successful use it directly,
# otherwise wrap as {"message": body} matching the normalize operator behavior.
# otherwise wrap as {"message": body} matching signozlogsnormalizer.
try:
json.loads(body)
self.body_v2 = body