Compare commits

...

20 Commits

Author SHA1 Message Date
Piyush Singariya
fabfe0e7d9 fix: integration tests 2026-03-12 23:32:38 +05:30
Piyush Singariya
c1e71ba199 fix: column change 2026-03-12 19:49:11 +05:30
Piyush Singariya
8b65c85a2a fix: column name fix 2026-03-12 17:53:06 +05:30
Piyush Singariya
d8c37e7a23 feat: string search tests 2026-03-12 17:42:44 +05:30
Piyush Singariya
b2f660bdcd Merge branch 'fix/message-keysearch' into json-qb-integration 2026-03-12 13:39:00 +05:30
Piyush Singariya
d9cad40d80 fix: key name in metadata 2026-03-12 13:07:39 +05:30
Piyush Singariya
8bd5ca8973 test: cleared 2026-03-12 13:02:25 +05:30
Piyush Singariya
fe94f6ad78 fix: excluding message from metadata fetch 2026-03-12 12:56:30 +05:30
Piyush Singariya
66e800aa2e fix: in progress 2026-03-11 20:56:24 +05:30
Piyush Singariya
3d6c0c6764 fix: test run success 2026-03-11 20:40:17 +05:30
Piyush Singariya
d9a1d64def feat: work in progress 2026-03-11 19:39:21 +05:30
Piyush Singariya
cb57b1fb00 Merge branch 'main' into json-qb-integration 2026-02-24 14:32:45 +05:30
Piyush Singariya
d5e2d140f3 Merge branch 'main' into json-qb-integration 2026-02-24 14:14:00 +05:30
Piyush Singariya
426d8e2cc8 Merge branch 'main' into json-qb-integration 2026-02-24 14:11:35 +05:30
Piyush Singariya
c0deb2a935 fix: tests 2026-02-02 13:48:18 +05:30
Piyush Singariya
e7923b757d Merge branch 'main' into json-qb-integration 2026-02-02 12:47:22 +05:30
Piyush Singariya
a7bc55bac9 revert: old json QB 2026-01-22 12:53:47 +05:30
Piyush Singariya
55ead8fb20 test: integration tests fixed 2026-01-21 22:34:22 +05:30
Piyush Singariya
0210f02397 Merge branch 'has-jsonqb' into json-qb-integration 2026-01-20 14:28:57 +05:30
Piyush Singariya
d5ac192dc4 feat: integration tests for new JSON QB 2026-01-19 12:22:46 +05:30
24 changed files with 2325 additions and 164 deletions

View File

@@ -18,13 +18,17 @@ type responseerroradditional struct {
func AsJSON(cause error) *JSON {
// See if this is an instance of the base error or not
_, c, m, _, u, a := Unwrapb(cause)
_, c, m, err, u, a := Unwrapb(cause)
rea := make([]responseerroradditional, len(a))
for k, v := range a {
rea[k] = responseerroradditional{v}
}
if err != nil {
rea = append(rea, responseerroradditional{err.Error()})
}
return &JSON{
Code: c.String(),
Message: m,

View File

@@ -658,7 +658,7 @@ func TestBaseRule_FilterNewSeries(t *testing.T) {
telemetryStore := telemetrystoretest.New(telemetrystore.Config{}, &queryMatcherAny{})
// Setup mock metadata store
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
// Create query parser
queryParser := queryparser.New(settings)

View File

@@ -33,7 +33,7 @@ func (c *conditionBuilder) conditionFor(
return "", err
}
if column.Type.GetType() == schema.ColumnTypeEnumJSON && querybuilder.BodyJSONQueryEnabled && key.FieldContext != telemetrytypes.FieldContextLog {
if column.Type.GetType() == schema.ColumnTypeEnumJSON && querybuilder.BodyJSONQueryEnabled && !key.Materialized {
valueType, value := InferDataType(value, operator, key)
cond, err := NewJSONConditionBuilder(key, valueType).buildJSONCondition(operator, value, sb)
if err != nil {
@@ -52,7 +52,7 @@ func (c *conditionBuilder) conditionFor(
}
// Check if this is a body JSON search - either by FieldContext
if key.FieldContext == telemetrytypes.FieldContextBody {
if key.FieldContext == telemetrytypes.FieldContextBody && !querybuilder.BodyJSONQueryEnabled {
tblFieldName, value = GetBodyJSONKey(ctx, key, operator, value)
}

View File

@@ -3,6 +3,7 @@ package telemetrylogs
import (
"fmt"
schema "github.com/SigNoz/signoz-otel-collector/cmd/signozschemamigrator/schema_migrator"
"github.com/SigNoz/signoz-otel-collector/constants"
"github.com/SigNoz/signoz/pkg/querybuilder"
qbtypes "github.com/SigNoz/signoz/pkg/types/querybuildertypes/querybuildertypesv5"
@@ -39,20 +40,22 @@ const (
BodyV2ColumnPrefix = constants.BodyV2ColumnPrefix
BodyPromotedColumnPrefix = constants.BodyPromotedColumnPrefix
MessageBodyField = "message"
MessageSubColumn = "body_v2.message"
bodySearchDefaultWarning = "body searches default to `body.message:string`. Use `body.<key>` to search a different field inside body"
)
var (
// mapping for body logical field to message sub column
// Here field context is log since message is a type hint and
// in a sense it is a direct column of log table and doesn't need any lambda expressions
// TODO(Piyush): Add description for detailed explanation of remapping of body to message
// BodyLogicalFieldJSONMapping is the canonical key for the "message" type hint.
// It lives under body (FieldContextBody) but Materialized=true signals the field
// mapper to access it as a direct sub-column (body_v2.message) rather than via
// dynamicElement(). Its name is the bare field name ("message"), not the column path.
BodyLogicalFieldJSONMapping = &telemetrytypes.TelemetryFieldKey{
Name: MessageSubColumn,
Name: MessageBodyField,
Signal: telemetrytypes.SignalLogs,
FieldContext: telemetrytypes.FieldContextLog,
FieldContext: telemetrytypes.FieldContextBody,
FieldDataType: telemetrytypes.FieldDataTypeString,
Materialized: true,
}
DefaultFullTextColumn = &telemetrytypes.TelemetryFieldKey{
Name: "body",
@@ -146,7 +149,15 @@ func enrichMapsForJSONBodyEnabled() {
if querybuilder.BodyJSONQueryEnabled {
DefaultFullTextColumn = BodyLogicalFieldJSONMapping
IntrinsicFields["body"] = *BodyLogicalFieldJSONMapping
// Register all key names that should resolve to the message type-hint column so
// QB can look them up directly: bare "message" and qualified "body_v2.message".
IntrinsicFields[MessageSubColumn] = *BodyLogicalFieldJSONMapping
IntrinsicFields[MessageBodyField] = *BodyLogicalFieldJSONMapping
logsV2Columns[MessageSubColumn] = &schema.Column{
Name: MessageSubColumn,
Type: schema.ColumnTypeString,
}
}
}

View File

@@ -58,7 +58,6 @@ var (
KeyType: schema.LowCardinalityColumnType{ElementType: schema.ColumnTypeString},
ValueType: schema.ColumnTypeString,
}},
MessageSubColumn: {Name: MessageSubColumn, Type: schema.ColumnTypeString},
}
)
@@ -89,6 +88,13 @@ func (m *fieldMapper) getColumn(_ context.Context, key *telemetrytypes.Telemetry
return logsV2Columns["attributes_bool"], nil
}
case telemetrytypes.FieldContextBody:
// Type hints (Materialized=true) have a direct physical sub-column in body_v2.
// Return a synthetic String column so the condition builder uses the direct path
// instead of the JSON condition builder (which expects a JSONPlan).
if key.Materialized {
// Type hints have a direct physical sub-column in body_v2 (e.g. body_v2.message).
return logsV2Columns[fmt.Sprintf("%s.%s", LogsV2BodyV2Column, key.Name)], nil
}
// Body context is for JSON body fields
// Use body_v2 if feature flag is enabled
if querybuilder.BodyJSONQueryEnabled {
@@ -249,34 +255,37 @@ func (m *fieldMapper) buildFieldForJSON(key *telemetrytypes.TelemetryFieldKey) (
node := plan[0]
expr := fmt.Sprintf("dynamicElement(%s, '%s')", node.FieldPath(), node.TerminalConfig.ElemType.StringValue())
if key.Materialized {
if len(plan) < 2 {
return "", errors.Newf(errors.TypeUnexpected, CodePromotedPlanMissing,
"plan length is less than 2 for promoted path: %s", key.Name)
}
// TODO(Piyush): Promoted path logic commented out. Materialized now means type hint
// promotion will be extracted from key field evolution
// (direct sub-column access), not a promoted body_promoted.* column.
// if key.Materialized {
// if len(plan) < 2 {
// return "", errors.Newf(errors.TypeUnexpected, CodePromotedPlanMissing,
// "plan length is less than 2 for promoted path: %s", key.Name)
// }
node := plan[1]
promotedExpr := fmt.Sprintf(
"dynamicElement(%s, '%s')",
node.FieldPath(),
node.TerminalConfig.ElemType.StringValue(),
)
// node := plan[1]
// promotedExpr := fmt.Sprintf(
// "dynamicElement(%s, '%s')",
// node.FieldPath(),
// node.TerminalConfig.ElemType.StringValue(),
// )
// dynamicElement returns NULL for scalar types or an empty array for array types.
if node.TerminalConfig.ElemType.IsArray {
expr = fmt.Sprintf(
"if(length(%s) > 0, %s, %s)",
promotedExpr,
promotedExpr,
expr,
)
} else {
// promoted column first then body_json column
// TODO(Piyush): Change this in future for better performance
expr = fmt.Sprintf("coalesce(%s, %s)", promotedExpr, expr)
}
// // dynamicElement returns NULL for scalar types or an empty array for array types.
// if node.TerminalConfig.ElemType.IsArray {
// expr = fmt.Sprintf(
// "if(length(%s) > 0, %s, %s)",
// promotedExpr,
// promotedExpr,
// expr,
// )
// } else {
// // promoted column first then body_json column
// // TODO(Piyush): Change this in future for better performance
// expr = fmt.Sprintf("coalesce(%s, %s)", promotedExpr, expr)
// }
}
// }
return expr, nil
}

View File

@@ -96,6 +96,9 @@ func TestStmtBuilderTimeSeriesBodyGroupByJSON(t *testing.T) {
}
}
/* Promoted path tests commented out — Materialized now means type hint (direct sub-column),
not a body_promoted.* column. These tests assumed the old coalesce(body_promoted.x, body_v2.x) path.
func TestStmtBuilderTimeSeriesBodyGroupByPromoted(t *testing.T) {
enable, disable := jsonQueryTestUtil(t)
enable()
@@ -156,6 +159,7 @@ func TestStmtBuilderTimeSeriesBodyGroupByPromoted(t *testing.T) {
})
}
}
*/
func TestStatementBuilderListQueryBodyHas(t *testing.T) {
enable, disable := jsonQueryTestUtil(t)
@@ -386,8 +390,8 @@ func TestStatementBuilderListQueryBody(t *testing.T) {
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{float64(2), float64(4), float64(2), float64(4), uint64(1747945619), uint64(1747983448), float64(2), float64(4), float64(2), float64(4), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) BETWEEN ? AND ?, arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), float64(2), float64(4), float64(2), float64(4), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
@@ -400,8 +404,8 @@ func TestStatementBuilderListQueryBody(t *testing.T) {
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{float64(2), float64(4), float64(2), float64(4), uint64(1747945619), uint64(1747983448), float64(2), float64(4), float64(2), float64(4), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (arrayExists(`body_v2.education`-> (arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')) OR arrayExists(`body_v2.education[].awards`-> toFloat64(dynamicElement(`body_v2.education[].awards`.`semester`, 'Int64')) IN (?, ?), arrayMap(x->dynamicElement(x, 'JSON'), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), float64(2), float64(4), float64(2), float64(4), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
@@ -525,6 +529,9 @@ func TestStatementBuilderListQueryBody(t *testing.T) {
}
}
/* Promoted path list-query tests commented out — Materialized now means type hint
(direct sub-column access), not a body_promoted.* column.
func TestStatementBuilderListQueryBodyPromoted(t *testing.T) {
enable, disable := jsonQueryTestUtil(t)
enable()
@@ -700,6 +707,7 @@ func TestStatementBuilderListQueryBodyPromoted(t *testing.T) {
})
}
}
*/
func TestStatementBuilderListQueryBodyMessage(t *testing.T) {
enable, disable := jsonQueryTestUtil(t)
@@ -707,19 +715,6 @@ func TestStatementBuilderListQueryBodyMessage(t *testing.T) {
defer disable()
statementBuilder := buildJSONTestStatementBuilder(t)
indexed := []*telemetrytypes.TelemetryFieldKey{
{
Name: "message",
Indexes: []telemetrytypes.JSONDataTypeIndex{
{
Type: telemetrytypes.String,
ColumnExpression: "body_promoted.message",
IndexExpression: "(lower(assumeNotNull(dynamicElement(body_promoted.message, 'String'))))",
},
},
},
}
testAddIndexedPaths(t, statementBuilder, indexed...)
cases := []struct {
name string
requestType qbtypes.RequestType
@@ -732,57 +727,57 @@ func TestStatementBuilderListQueryBodyMessage(t *testing.T) {
requestType: qbtypes.RequestTypeRaw,
query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
Signal: telemetrytypes.SignalLogs,
Filter: &qbtypes.Filter{Expression: "body.message Exists"},
Filter: &qbtypes.Filter{Expression: "message Exists"},
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (dynamicElement(body_v2.`message`, 'String') IS NOT NULL) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
{
name: "body.message equals to empty string",
requestType: qbtypes.RequestTypeRaw,
query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
Signal: telemetrytypes.SignalLogs,
Filter: &qbtypes.Filter{Expression: "body.message = ''"},
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (dynamicElement(body_v2.`message`, 'String') = ?) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND body_v2.message <> ? AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), "", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
{
name: "body.message equals to 'Iron Award'",
requestType: qbtypes.RequestTypeRaw,
query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
Signal: telemetrytypes.SignalLogs,
Filter: &qbtypes.Filter{Expression: "body.message = 'Iron Award'"},
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (dynamicElement(body_v2.`message`, 'String') = ?) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), "Iron Award", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
{
name: "body.message contains 'Iron Award'",
requestType: qbtypes.RequestTypeRaw,
query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
Signal: telemetrytypes.SignalLogs,
Filter: &qbtypes.Filter{Expression: "body.message Contains 'Iron Award'"},
Limit: 10,
},
expected: qbtypes.Statement{
Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (LOWER(dynamicElement(body_v2.`message`, 'String')) LIKE LOWER(?)) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
Args: []any{uint64(1747945619), uint64(1747983448), "%Iron Award%", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
},
expectedErr: nil,
},
// {
// name: "body.message equals to empty string",
// requestType: qbtypes.RequestTypeRaw,
// query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
// Signal: telemetrytypes.SignalLogs,
// Filter: &qbtypes.Filter{Expression: "body.message = ''"},
// Limit: 10,
// },
// expected: qbtypes.Statement{
// Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND body_v2.message = ? AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
// Args: []any{uint64(1747945619), uint64(1747983448), "", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
// },
// expectedErr: nil,
// },
// {
// name: "body.message equals to 'Iron Award'",
// requestType: qbtypes.RequestTypeRaw,
// query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
// Signal: telemetrytypes.SignalLogs,
// Filter: &qbtypes.Filter{Expression: "body.message = 'Iron Award'"},
// Limit: 10,
// },
// expected: qbtypes.Statement{
// Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND body_v2.message = ? AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
// Args: []any{uint64(1747945619), uint64(1747983448), "Iron Award", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
// },
// expectedErr: nil,
// },
// {
// name: "message contains 'Iron Award'",
// requestType: qbtypes.RequestTypeRaw,
// query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
// Signal: telemetrytypes.SignalLogs,
// Filter: &qbtypes.Filter{Expression: "message Contains 'Iron Award'"},
// Limit: 10,
// },
// expected: qbtypes.Statement{
// Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND LOWER(body_v2.message) LIKE LOWER(?) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
// Args: []any{uint64(1747945619), uint64(1747983448), "%Iron Award%", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
// },
// expectedErr: nil,
// },
}
for _, c := range cases {
@@ -804,8 +799,7 @@ func TestStatementBuilderListQueryBodyMessage(t *testing.T) {
}
func buildTestTelemetryMetadataStore(t *testing.T, promotedPaths ...string) *telemetrytypestest.MockMetadataStore {
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(IntrinsicFields)
types, _ := telemetrytypes.TestJSONTypeSet()
for path, jsonTypes := range types {
promoted := false
@@ -842,11 +836,14 @@ func buildJSONTestStatementBuilder(t *testing.T, promotedPaths ...string) *logQu
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
resourceFilterFM := resourcefilter.NewFieldMapper()
resourceFilterCB := resourcefilter.NewConditionBuilder(resourceFilterFM)
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
resourceFilterStmtBuilder := resourcefilter.NewLogResourceFilterStatementBuilder(
instrumentationtest.New().ToProviderSettings(),
fm,
cb,
resourceFilterFM,
resourceFilterCB,
mockMetadataStore,
DefaultFullTextColumn,
GetBodyJSONKey,
@@ -866,19 +863,6 @@ func buildJSONTestStatementBuilder(t *testing.T, promotedPaths ...string) *logQu
return statementBuilder
}
func testAddIndexedPaths(t *testing.T, statementBuilder *logQueryStatementBuilder, telemetryFieldKeys ...*telemetrytypes.TelemetryFieldKey) {
mockMetadataStore := statementBuilder.metadataStore.(*telemetrytypestest.MockMetadataStore)
for _, key := range telemetryFieldKeys {
if strings.Contains(key.Name, telemetrytypes.ArraySep) || strings.Contains(key.Name, telemetrytypes.ArrayAnyIndex) {
t.Fatalf("array paths are not supported: %s", key.Name)
}
for _, storedKey := range mockMetadataStore.KeysMap[key.Name] {
storedKey.Indexes = append(storedKey.Indexes, key.Indexes...)
}
}
}
func jsonQueryTestUtil(_ *testing.T) (func(), func()) {
querybuilder.BodyJSONQueryEnabled = true
base := telemetrytypes.TelemetryFieldKey{
@@ -897,6 +881,8 @@ func jsonQueryTestUtil(_ *testing.T) (func(), func()) {
DefaultFullTextColumn = &base
IntrinsicFields["body"] = base
delete(IntrinsicFields, MessageSubColumn)
delete(IntrinsicFields, MessageBodyField)
delete(logsV2Columns, MessageSubColumn)
}
return enable, disable

View File

@@ -18,7 +18,7 @@ import (
func resourceFilterStmtBuilder() qbtypes.StatementBuilder[qbtypes.LogAggregation] {
fm := resourcefilter.NewFieldMapper()
cb := resourcefilter.NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
keysMap := buildCompleteFieldKeyMap()
for _, keys := range keysMap {
for _, key := range keys {
@@ -196,7 +196,7 @@ func TestStatementBuilderTimeSeries(t *testing.T) {
},
}
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
@@ -316,7 +316,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
},
}
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
@@ -456,7 +456,7 @@ func TestStatementBuilderListQueryResourceTests(t *testing.T) {
},
}
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
@@ -532,7 +532,7 @@ func TestStatementBuilderTimeSeriesBodyGroupBy(t *testing.T) {
},
}
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
@@ -627,7 +627,7 @@ func TestStatementBuilderListQueryServiceCollision(t *testing.T) {
},
}
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMapCollision()
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
@@ -850,7 +850,7 @@ func TestAdjustKey(t *testing.T) {
}
fm := NewFieldMapper()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMapCollision()
cb := NewConditionBuilder(fm)
@@ -1005,9 +1005,7 @@ func TestStmtBuilderBodyField(t *testing.T) {
disable()
}
// build the key map after enabling/disabling body JSON query
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(IntrinsicFields)
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
resourceFilterStmtBuilder := resourceFilterStmtBuilder()
statementBuilder := NewLogQueryStatementBuilder(

View File

@@ -931,6 +931,13 @@ func buildCompleteFieldKeyMap() map[string][]*telemetrytypes.TelemetryFieldKey {
Materialized: true,
},
},
"body": {
{
Name: "body",
FieldContext: telemetrytypes.FieldContextLog,
FieldDataType: telemetrytypes.FieldDataTypeString,
},
},
}
for _, keys := range keysMap {
@@ -939,10 +946,6 @@ func buildCompleteFieldKeyMap() map[string][]*telemetrytypes.TelemetryFieldKey {
}
}
// add intrinsic fields to the map
for fieldName, key := range IntrinsicFields {
keysMap[fieldName] = append(keysMap[fieldName], &key)
}
return keysMap
}

View File

@@ -54,6 +54,7 @@ func (t *telemetryMetaStore) fetchBodyJSONPaths(ctx context.Context,
instrumentationtypes.CodeNamespace: "metadata",
instrumentationtypes.CodeFunctionName: "fetchBodyJSONPaths",
})
query, args, limit := buildGetBodyJSONPathsQuery(fieldKeySelectors)
rows, err := t.telemetrystore.ClickhouseDB().Query(ctx, query, args...)
if err != nil {
@@ -184,6 +185,9 @@ func buildGetBodyJSONPathsQuery(fieldKeySelectors []*telemetrytypes.FieldKeySele
limit += fieldKeySelector.Limit
}
sb.Where(sb.Or(orClauses...))
// mesasge field is skipped; since it is a type hint and is handled by the field mapper
// TODO(Piyush): If typehints increases in future, use aftership parser to skip type hints here
sb.Where(sb.NotEqual("path", telemetrylogs.MessageBodyField))
// Group by path to get unique paths with aggregated types
sb.GroupBy("path")

View File

@@ -527,7 +527,7 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
if err != nil {
return nil, false, errors.Wrap(err, errors.TypeInternal, errors.CodeInternal, ErrFailedToGetLogsKeys.Error())
}
key, ok := setOfKeys[name+";"+fieldContext.StringValue()+";"+fieldDataType.StringValue()]
key, ok := setOfKeys[fieldContext.StringValue()+"."+name+":"+fieldDataType.StringValue()]
// if there is no materialised column, create a key with the field context and data type
if !ok {
@@ -540,7 +540,7 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
}
mapOfKeys[key.Name] = append(mapOfKeys[key.Name], key)
setOfKeys[name+";"+fieldContext.StringValue()+";"+fieldDataType.StringValue()] = key
setOfKeys[key.Text()] = key
}
if rows.Err() != nil {
@@ -566,12 +566,15 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
if found {
if field, exists := telemetrylogs.IntrinsicFields[key]; exists {
if _, added := setOfKeys[field.Name+";"+field.FieldContext.StringValue()+";"+field.FieldDataType.StringValue()]; !added {
mapOfKeys[field.Name] = append(mapOfKeys[field.Name], &field)
// intrinsic field can be a logical field mapped to a different physical location
mapOfKeys[key] = append(mapOfKeys[key], &field)
// Register by physical name only once and only when it differs from the
// logical key — if they are the same, the always-register below covers it.
if _, added := setOfKeys[field.Text()]; !added {
mapOfKeys[field.Text()] = append(mapOfKeys[field.Text()], &field)
}
continue
// Always register the logical key so that every alias in IntrinsicFields
// (e.g. "message", "body_v2.message") independently resolves to the same
// physical field in the keys map.
mapOfKeys[key] = append(mapOfKeys[key], &field)
}
}
}

View File

@@ -159,7 +159,7 @@ func TestStatementBuilder(t *testing.T) {
fm := telemetrymetrics.NewFieldMapper()
cb := telemetrymetrics.NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
keys, err := telemetrytypestest.LoadFieldKeysFromJSON("testdata/keys_map.json")
if err != nil {
t.Fatalf("failed to load field keys: %v", err)

View File

@@ -221,7 +221,7 @@ func TestStatementBuilder(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
keys, err := telemetrytypestest.LoadFieldKeysFromJSON("testdata/keys_map.json")
if err != nil {
t.Fatalf("failed to load field keys: %v", err)

View File

@@ -18,7 +18,7 @@ import (
func resourceFilterStmtBuilder() qbtypes.StatementBuilder[qbtypes.TraceAggregation] {
fm := resourcefilter.NewFieldMapper()
cb := resourcefilter.NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
return resourcefilter.NewTraceResourceFilterStatementBuilder(
@@ -368,7 +368,7 @@ func TestStatementBuilder(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
@@ -664,7 +664,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
@@ -771,7 +771,7 @@ func TestStatementBuilderListQueryWithCorruptData(t *testing.T) {
t.Run(c.name, func(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = c.keysMap
if mockMetadataStore.KeysMap == nil {
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
@@ -927,7 +927,7 @@ func TestStatementBuilderTraceQuery(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
@@ -1145,7 +1145,7 @@ func TestAdjustKey(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
resourceFilterStmtBuilder := resourceFilterStmtBuilder()
@@ -1420,7 +1420,7 @@ func TestAdjustKeys(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
resourceFilterStmtBuilder := resourceFilterStmtBuilder()

View File

@@ -388,7 +388,7 @@ func TestTraceOperatorStatementBuilder(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)
@@ -504,7 +504,7 @@ func TestTraceOperatorStatementBuilderErrors(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
aggExprRewriter := querybuilder.NewAggExprRewriter(instrumentationtest.New().ToProviderSettings(), nil, fm, cb, nil)

View File

@@ -19,7 +19,7 @@ func TestTraceTimeRangeOptimization(t *testing.T) {
fm := NewFieldMapper()
cb := NewConditionBuilder(fm)
mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
mockMetadataStore := telemetrytypestest.NewMockMetadataStore(nil)
mockMetadataStore.KeysMap = buildCompleteFieldKeyMap()
mockMetadataStore.KeysMap["trace_id"] = []*telemetrytypes.TelemetryFieldKey{{

View File

@@ -40,7 +40,7 @@ type TelemetryFieldKey struct {
JSONDataType *JSONDataType `json:"-"`
JSONPlan JSONAccessPlan `json:"-"`
Indexes []JSONDataTypeIndex `json:"-"`
Materialized bool `json:"-"` // refers to promoted in case of body.... fields
Materialized bool `json:"-"` // refers to type hint in case of JSON column fields
}
func (f *TelemetryFieldKey) KeyNameContainsArray() bool {

View File

@@ -21,10 +21,20 @@ type MockMetadataStore struct {
PromotedPathsMap map[string]bool
LogsJSONIndexesMap map[string][]schemamigrator.Index
LookupKeysMap map[telemetrytypes.MetricMetadataLookupKey]int64
// StaticFields holds signal-specific intrinsic field definitions (e.g. telemetrylogs.IntrinsicFields).
// It is injected into GetKeys / GetKey results, mirroring what the real metadata store does when
// it reads telemetrylogs.IntrinsicFields directly. Callers pass their package's IntrinsicFields
// map at construction time to avoid a circular import.
StaticFields map[string]telemetrytypes.TelemetryFieldKey
}
// NewMockMetadataStore creates a new instance of MockMetadataStore with initialized maps
func NewMockMetadataStore() *MockMetadataStore {
// NewMockMetadataStore creates a new instance of MockMetadataStore with initialized maps.
// Pass the signal-specific intrinsic fields (e.g. telemetrylogs.IntrinsicFields) so the mock
// mirrors what the real metadata store does when injecting those definitions into key results.
func NewMockMetadataStore(intrinsicFields map[string]telemetrytypes.TelemetryFieldKey) *MockMetadataStore {
if intrinsicFields == nil {
intrinsicFields = make(map[string]telemetrytypes.TelemetryFieldKey)
}
return &MockMetadataStore{
KeysMap: make(map[string][]*telemetrytypes.TelemetryFieldKey),
RelatedValuesMap: make(map[string][]string),
@@ -34,12 +44,13 @@ func NewMockMetadataStore() *MockMetadataStore {
PromotedPathsMap: make(map[string]bool),
LogsJSONIndexesMap: make(map[string][]schemamigrator.Index),
LookupKeysMap: make(map[telemetrytypes.MetricMetadataLookupKey]int64),
StaticFields: intrinsicFields,
}
}
// GetKeys returns a map of field keys types.TelemetryFieldKey by name
func (m *MockMetadataStore) GetKeys(ctx context.Context, fieldKeySelector *telemetrytypes.FieldKeySelector) (map[string][]*telemetrytypes.TelemetryFieldKey, bool, error) {
setOfKeys := make(map[string]*telemetrytypes.TelemetryFieldKey)
result := make(map[string][]*telemetrytypes.TelemetryFieldKey)
// If selector is nil, return all keys
@@ -47,22 +58,42 @@ func (m *MockMetadataStore) GetKeys(ctx context.Context, fieldKeySelector *telem
return m.KeysMap, true, nil
}
// Apply selector logic
// Apply selector logic from KeysMap
for name, keys := range m.KeysMap {
// Check if name matches
if matchesName(fieldKeySelector, name) {
filteredKeys := []*telemetrytypes.TelemetryFieldKey{}
for _, key := range keys {
if matchesKey(fieldKeySelector, key) {
filteredKeys = append(filteredKeys, key)
if _, exists := setOfKeys[key.Text()]; !exists {
result[name] = append(result[name], key)
setOfKeys[key.Text()] = key
}
}
}
if len(filteredKeys) > 0 {
result[name] = filteredKeys
}
}
}
// Inject StaticFields (e.g. IntrinsicFields), mirroring the real metadata store.
// StaticFields take precedence over KeysMap entries for the same logical key.
// Each logical key always gets its own entry (so "message" and "body.message" both
// resolve to the IntrinsicField independently). The physical name is registered
// only once to avoid duplicate entries in that slot.
for key, field := range m.StaticFields {
if !matchesName(fieldKeySelector, key) {
continue
}
// Register by physical name only once and only when it differs from the
// logical key — if they are the same, the always-register below covers it.
if _, exists := setOfKeys[field.Text()]; !exists {
result[field.Text()] = append(result[field.Text()], &field)
setOfKeys[field.Text()] = &field
}
// Always register the logical key so that every alias in IntrinsicFields
// (e.g. "message", "body_v2.message") independently resolves to the same
// physical field in the keys map.
result[key] = append(result[key], &field)
}
return result, true, nil
}
@@ -109,7 +140,7 @@ func (m *MockMetadataStore) GetKey(ctx context.Context, fieldKeySelector *teleme
result := []*telemetrytypes.TelemetryFieldKey{}
// Find keys matching the selector
// Find keys matching the selector from KeysMap
for name, keys := range m.KeysMap {
if matchesName(fieldKeySelector, name) {
for _, key := range keys {
@@ -120,6 +151,14 @@ func (m *MockMetadataStore) GetKey(ctx context.Context, fieldKeySelector *teleme
}
}
// Add matching StaticFields (e.g. IntrinsicFields), same as the real metadata store does
for key, field := range m.StaticFields {
if fieldKeySelector.Name == "" || strings.Contains(key, fieldKeySelector.Name) {
fieldCopy := field
result = append(result, &fieldCopy)
}
}
return result, nil
}
@@ -291,7 +330,7 @@ func (m *MockMetadataStore) FetchTemporalityMulti(ctx context.Context, queryTime
return result, nil
}
// FetchTemporalityMulti fetches the temporality for multiple metrics
// FetchTemporalityAndTypeMulti fetches the temporality and type for multiple metrics
func (m *MockMetadataStore) FetchTemporalityAndTypeMulti(ctx context.Context, queryTimeRangeStartTs, queryTimeRangeEndTs uint64, metricNames ...string) (map[string]metrictypes.Temporality, map[string]metrictypes.Type, error) {
temporalities := make(map[string]metrictypes.Temporality)
types := make(map[string]metrictypes.Type)

View File

@@ -64,8 +64,7 @@ func TestJSONTypeSet() (map[string][]JSONDataType, MetadataStore) {
"interests[].entities[].reviews[].entries[].metadata[].positions[].duration": {Int64, Float64},
"interests[].entities[].reviews[].entries[].metadata[].positions[].unit": {String},
"interests[].entities[].reviews[].entries[].metadata[].positions[].ratings": {ArrayInt64, ArrayString},
"message": {String},
"tags": {ArrayString},
"tags": {ArrayString},
}
return types, nil

View File

@@ -1,5 +1,19 @@
import json
import os
import tempfile
import pytest
if not os.environ.get("DOCKER_CONFIG"):
os.environ["DOCKER_CONFIG"] = tempfile.mkdtemp(prefix="docker-config-")
os.environ.setdefault("DOCKER_CREDENTIAL_STORE", "")
os.environ.setdefault("DOCKER_CREDENTIAL_HELPER", "")
docker_config_path = os.path.join(os.environ["DOCKER_CONFIG"], "config.json")
if not os.path.exists(docker_config_path):
with open(docker_config_path, "w", encoding="utf-8") as config_file:
json.dump({"auths": {}, "credsStore": ""}, config_file)
pytest_plugins = [
"fixtures.auth",
"fixtures.clickhouse",
@@ -22,6 +36,7 @@ pytest_plugins = [
"fixtures.notification_channel",
"fixtures.alerts",
"fixtures.cloudintegrations",
"fixtures.jsontypeexporter",
]
@@ -59,7 +74,7 @@ def pytest_addoption(parser: pytest.Parser):
parser.addoption(
"--clickhouse-version",
action="store",
default="25.5.6",
default="25.8.6",
help="clickhouse version",
)
parser.addoption(
@@ -74,3 +89,5 @@ def pytest_addoption(parser: pytest.Parser):
default="v0.129.7",
help="schema migrator version",
)

View File

@@ -0,0 +1,421 @@
"""
Simpler version of jsontypeexporter for test fixtures.
This exports JSON type metadata to the path_types table by parsing JSON bodies
and extracting all paths with their types, similar to how the real jsontypeexporter works.
"""
import datetime
import json
from abc import ABC
from typing import TYPE_CHECKING, Any, Callable, Dict, Generator, List, Optional, Set, Union
import numpy as np
import pytest
from fixtures import types
if TYPE_CHECKING:
from fixtures.logs import Logs
class JSONPathType(ABC):
"""Represents a JSON path with its type information"""
path: str
type: str
last_seen: np.uint64
def __init__(
self,
path: str,
type: str, # pylint: disable=redefined-builtin
last_seen: Optional[datetime.datetime] = None,
) -> None:
self.path = path
self.type = type
if last_seen is None:
last_seen = datetime.datetime.now()
self.last_seen = np.uint64(int(last_seen.timestamp() * 1e9))
def np_arr(self) -> np.array:
"""Return path type data as numpy array for database insertion"""
return np.array([self.path, self.type, self.last_seen])
# Constants matching jsontypeexporter
ARRAY_SEPARATOR = "[]." # Used in paths like "education[].name"
ARRAY_SUFFIX = "[]" # Used when traversing into array element objects
def _infer_array_type_from_type_strings(types: List[str]) -> Optional[str]:
"""
Infer array type from a list of pre-classified type strings.
Matches jsontypeexporter's inferArrayMask logic (v0.144.2+).
Type strings are: "JSON", "String", "Bool", "Float64", "Int64"
SuperTyping rules (matching Go inferArrayMask):
- JSON alone → Array(JSON)
- JSON + any primitive → Array(Dynamic)
- String alone → Array(Nullable(String)); String + other → Array(Dynamic)
- Float64 wins over Int64 and Bool
- Int64 wins over Bool
- Bool alone → Array(Nullable(Bool))
"""
if len(types) == 0:
return None
unique = set(types)
has_json = "JSON" in unique
# hasPrimitive mirrors Go: (hasJSON && len(unique) > 1) || (!hasJSON && len(unique) > 0)
has_primitive = (has_json and len(unique) > 1) or (not has_json and len(unique) > 0)
if has_json:
if not has_primitive:
return "Array(JSON)"
return "Array(Dynamic)"
# ---- Primitive Type Resolution (Float > Int > Bool) ----
if "String" in unique:
if len(unique) > 1:
return "Array(Dynamic)"
return "Array(Nullable(String))"
if "Float64" in unique:
return "Array(Nullable(Float64))"
if "Int64" in unique:
return "Array(Nullable(Int64))"
if "Bool" in unique:
return "Array(Nullable(Bool))"
return "Array(Dynamic)"
def _infer_array_type(elements: List[Any]) -> Optional[str]:
"""
Infer array type from raw Python list elements.
Classifies each element then delegates to _infer_array_type_from_type_strings.
"""
if len(elements) == 0:
return None
types = []
for elem in elements:
if elem is None:
continue
if isinstance(elem, dict):
types.append("JSON")
elif isinstance(elem, str):
types.append("String")
elif isinstance(elem, bool): # must be before int (bool is subclass of int)
types.append("Bool")
elif isinstance(elem, float):
types.append("Float64")
elif isinstance(elem, int):
types.append("Int64")
return _infer_array_type_from_type_strings(types)
def _python_type_to_clickhouse_type(value: Any) -> str:
"""
Convert Python type to ClickHouse JSON type string.
Maps Python types to ClickHouse JSON data types.
"""
if value is None:
return "String" # Default for null values
if isinstance(value, bool):
return "Bool"
elif isinstance(value, int):
return "Int64"
elif isinstance(value, float):
return "Float64"
elif isinstance(value, str):
return "String"
elif isinstance(value, list):
# Use the sophisticated array type inference
array_type = _infer_array_type(value)
return array_type if array_type else "Array(Dynamic)"
elif isinstance(value, dict):
return "JSON"
else:
return "String" # Default fallback
def _extract_json_paths(
obj: Any,
current_path: str = "",
path_types: Optional[Dict[str, Set[str]]] = None,
level: int = 0,
) -> Dict[str, Set[str]]:
"""
Recursively extract all paths and their types from a JSON object.
Matches jsontypeexporter's analyzePValue logic.
Args:
obj: The JSON object to traverse
current_path: Current path being built (e.g., "user.name")
path_types: Dictionary mapping paths to sets of types found
level: Current nesting level (for depth limiting)
Returns:
Dictionary mapping paths to sets of type strings
"""
if path_types is None:
path_types = {}
if obj is None:
if current_path:
if current_path not in path_types:
path_types[current_path] = set()
path_types[current_path].add("String") # Null defaults to String
return path_types
if isinstance(obj, dict):
# For objects, add the object itself and recurse into keys
if current_path:
if current_path not in path_types:
path_types[current_path] = set()
path_types[current_path].add("JSON")
for key, value in obj.items():
# Build the path for this key
if current_path:
new_path = f"{current_path}.{key}"
else:
new_path = key
# Recurse into the value
_extract_json_paths(value, new_path, path_types, level + 1)
elif isinstance(obj, list):
# Skip empty arrays
if len(obj) == 0:
return path_types
# Collect types from array elements (matching Go: types := make([]pcommon.ValueType, 0, s.Len()))
types = []
for item in obj:
if isinstance(item, dict):
# When traversing into array element objects, use ArraySuffix ([])
# This matches: prefix+ArraySuffix in the Go code
# Example: if current_path is "education", we use "education[]" to traverse into objects
array_prefix = current_path + ARRAY_SUFFIX if current_path else ""
for key, value in item.items():
if array_prefix:
# Use array separator: education[].name
array_path = f"{array_prefix}.{key}"
else:
array_path = key
# Recurse without increasing level (matching Go behavior)
_extract_json_paths(value, array_path, path_types, level)
types.append("JSON")
elif isinstance(item, list):
# Arrays inside arrays are not supported - skip the whole path
# Matching Go: e.logger.Error("arrays inside arrays are not supported!", ...); return nil
return path_types
elif isinstance(item, str):
types.append("String")
elif isinstance(item, bool):
types.append("Bool")
elif isinstance(item, float):
types.append("Float64")
elif isinstance(item, int):
types.append("Int64")
# Infer array type from collected types (matching Go: if mask := inferArrayMask(types); mask != 0)
if len(types) > 0:
array_type = _infer_array_type_from_type_strings(types)
if array_type and current_path:
if current_path not in path_types:
path_types[current_path] = set()
path_types[current_path].add(array_type)
else:
# Primitive value (string, number, bool)
if current_path:
if current_path not in path_types:
path_types[current_path] = set()
obj_type = _python_type_to_clickhouse_type(obj)
path_types[current_path].add(obj_type)
return path_types
def _parse_json_bodies_and_extract_paths(
json_bodies: List[str],
timestamp: Optional[datetime.datetime] = None,
) -> List[JSONPathType]:
"""
Parse JSON bodies and extract all paths with their types.
This mimics the behavior of jsontypeexporter.
Args:
json_bodies: List of JSON body strings to parse
timestamp: Timestamp to use for last_seen (defaults to now)
Returns:
List of JSONPathType objects with all discovered paths and types
"""
if timestamp is None:
timestamp = datetime.datetime.now()
# Aggregate all paths and their types across all JSON bodies
all_path_types: Dict[str, Set[str]] = {}
for json_body in json_bodies:
try:
parsed = json.loads(json_body)
_extract_json_paths(parsed, "", all_path_types, level=0)
except (json.JSONDecodeError, TypeError):
# Skip invalid JSON
continue
# Convert to list of JSONPathType objects
# Each path can have multiple types, so we create one JSONPathType per type
path_type_objects: List[JSONPathType] = []
for path, types_set in all_path_types.items():
for type_str in types_set:
path_type_objects.append(
JSONPathType(path=path, type=type_str, last_seen=timestamp)
)
return path_type_objects
@pytest.fixture(name="export_json_types", scope="function")
def export_json_types(
clickhouse: types.TestContainerClickhouse,
request: pytest.FixtureRequest, # To access migrator fixture
) -> Generator[Callable[[Union[List[JSONPathType], List[str], List[Any]]], None], Any, None]:
"""
Fixture for exporting JSON type metadata to the path_types table.
This is a simpler version of jsontypeexporter for test fixtures.
The function can accept:
1. List of JSONPathType objects (manual specification)
2. List of JSON body strings (auto-extract paths)
3. List of Logs objects (extract from body_json field)
Usage examples:
# Manual specification
export_json_types([
JSONPathType(path="user.name", type="String"),
JSONPathType(path="user.age", type="Int64"),
])
# Auto-extract from JSON strings
export_json_types([
'{"user": {"name": "alice", "age": 25}}',
'{"user": {"name": "bob", "age": 30}}',
])
# Auto-extract from Logs objects
export_json_types(logs_list)
"""
# Ensure migrator has run to create the table
try:
request.getfixturevalue("migrator")
except Exception:
# If migrator fixture is not available, that's okay - table might already exist
pass
def _export_json_types(
data: Union[List[JSONPathType], List[str], List[Any]], # List[Logs] but avoiding circular import
) -> None:
"""
Export JSON type metadata to signoz_metadata.distributed_json_path_types table.
This table stores path and type information for body JSON fields.
"""
path_types: List[JSONPathType] = []
if len(data) == 0:
return
# Determine input type and convert to JSONPathType list
first_item = data[0]
if isinstance(first_item, JSONPathType):
# Already JSONPathType objects
path_types = data # type: ignore
elif isinstance(first_item, str):
# List of JSON strings - parse and extract paths
path_types = _parse_json_bodies_and_extract_paths(data) # type: ignore
else:
# Assume it's a list of Logs objects - extract body_v2
json_bodies: List[str] = []
for log in data: # type: ignore
# Try to get body_v2 attribute
if hasattr(log, "body_v2") and log.body_v2:
json_bodies.append(log.body_v2)
elif hasattr(log, "body") and log.body:
# Fallback to body if body_v2 not available
try:
# Try to parse as JSON
json.loads(log.body)
json_bodies.append(log.body)
except (json.JSONDecodeError, TypeError):
pass
if json_bodies:
path_types = _parse_json_bodies_and_extract_paths(json_bodies)
if len(path_types) == 0:
return
clickhouse.conn.insert(
database="signoz_metadata",
table="distributed_json_path_types",
data=[path_type.np_arr() for path_type in path_types],
column_names=[
"path",
"type",
"last_seen",
],
)
yield _export_json_types
# Cleanup - truncate the local table after tests (following pattern from logs fixture)
clickhouse.conn.query(
f"TRUNCATE TABLE signoz_metadata.json_path_types ON CLUSTER '{clickhouse.env['SIGNOZ_TELEMETRYSTORE_CLICKHOUSE_CLUSTER']}' SYNC"
)
@pytest.fixture(name="export_promoted_paths", scope="function")
def export_promoted_paths(
clickhouse: types.TestContainerClickhouse,
request: pytest.FixtureRequest, # To access migrator fixture
) -> Generator[Callable[[List[str]], None], Any, None]:
"""
Fixture for exporting promoted JSON paths to the promoted paths table.
"""
# Ensure migrator has run to create the table
try:
request.getfixturevalue("migrator")
except Exception:
# If migrator fixture is not available, that's okay - table might already exist
pass
def _export_promoted_paths(paths: List[str]) -> None:
if len(paths) == 0:
return
now_ms = int(datetime.datetime.now().timestamp() * 1000)
rows = [(path, now_ms) for path in paths]
clickhouse.conn.insert(
database="signoz_metadata",
table="distributed_json_promoted_paths",
data=rows,
column_names=[
"path",
"created_at",
],
)
yield _export_promoted_paths
clickhouse.conn.query(
f"TRUNCATE TABLE signoz_metadata.json_promoted_paths ON CLUSTER '{clickhouse.env['SIGNOZ_TELEMETRYSTORE_CLICKHOUSE_CLUSTER']}' SYNC"
)

View File

@@ -100,6 +100,8 @@ class Logs(ABC):
severity_text: str
severity_number: np.uint8
body: str
body_v2: str
body_promoted: str
attributes_string: dict[str, str]
attributes_number: dict[str, np.float64]
attributes_bool: dict[str, bool]
@@ -119,6 +121,8 @@ class Logs(ABC):
resources: dict[str, Any] = {},
attributes: dict[str, Any] = {},
body: str = "default body",
body_v2: Optional[str] = None,
body_promoted: Optional[str] = None,
severity_text: str = "INFO",
trace_id: str = "",
span_id: str = "",
@@ -163,6 +167,33 @@ class Logs(ABC):
# Set body
self.body = body
# Set body_v2 - if body is JSON, parse and stringify it, otherwise use empty string
# ClickHouse accepts String input for JSON column
if body_v2 is not None:
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.
try:
json.loads(body)
self.body_v2 = body
except (json.JSONDecodeError, TypeError):
self.body_v2 = json.dumps({"message": body})
# Set body_promoted - must be valid JSON
# Tests will explicitly pass promoted column's content, but we validate it
if body_promoted is not None:
# Validate that it's valid JSON
try:
json.loads(body_promoted)
self.body_promoted = body_promoted
except (json.JSONDecodeError, TypeError):
# If invalid, default to empty JSON object
self.body_promoted = "{}"
else:
# Default to empty JSON object (valid JSON)
self.body_promoted = "{}"
# Process resources and attributes
self.resources_string = {k: str(v) for k, v in resources.items()}
for k, v in self.resources_string.items():
@@ -319,6 +350,8 @@ class Logs(ABC):
self.severity_text,
self.severity_number,
self.body,
self.body_v2,
self.body_promoted,
self.attributes_string,
self.attributes_number,
self.attributes_bool,
@@ -463,6 +496,8 @@ def insert_logs(
"severity_text",
"severity_number",
"body",
"body_v2",
"body_promoted",
"attributes_string",
"attributes_number",
"attributes_bool",

View File

@@ -20,12 +20,14 @@ def migrator(
"""
def create() -> None:
version = request.config.getoption("--schema-migrator-version")
# Hardcode version for new QB tests
version = "v0.144.3-rc.1"
client = docker.from_env()
container = client.containers.run(
image=f"signoz/signoz-schema-migrator:{version}",
command=f"sync --replication=true --cluster-name=cluster --up= --dsn={clickhouse.env["SIGNOZ_TELEMETRYSTORE_CLICKHOUSE_DSN"]}",
environment={"ENABLE_LOGS_MIGRATIONS_V2": "1"},
detach=True,
auto_remove=False,
network=network.id,
@@ -44,6 +46,7 @@ def migrator(
container = client.containers.run(
image=f"signoz/signoz-schema-migrator:{version}",
command=f"async --replication=true --cluster-name=cluster --up= --dsn={clickhouse.env["SIGNOZ_TELEMETRYSTORE_CLICKHOUSE_DSN"]}",
environment={"ENABLE_LOGS_MIGRATIONS_V2": "1"},
detach=True,
auto_remove=False,
network=network.id,

View File

@@ -73,6 +73,7 @@ def signoz( # pylint: disable=too-many-arguments,too-many-positional-arguments
"SIGNOZ_ALERTMANAGER_SIGNOZ_POLL__INTERVAL": "5s",
"SIGNOZ_ALERTMANAGER_SIGNOZ_ROUTE_GROUP__WAIT": "1s",
"SIGNOZ_ALERTMANAGER_SIGNOZ_ROUTE_GROUP__INTERVAL": "5s",
"BODY_JSON_QUERY_ENABLED": "true",
}
| sqlstore.env
| clickhouse.env

File diff suppressed because it is too large Load Diff