feat: added span_count using the MV and index v3 traces table

deploy/docker-swarm/clickhouse-setup/docker-compose.yaml

@@ -150,8 +150,7 @@ services:
     command:
       [
         "-config=/root/config/prometheus.yml",
-        "--use-logs-new-schema=true",
-        "--use-trace-new-schema=true"
+        "--use-logs-new-schema=true"
       ]
     # ports:
     #   - "6060:6060"     # pprof port

deploy/docker-swarm/clickhouse-setup/otel-collector-config.yaml

@@ -110,7 +110,6 @@ exporters:
   clickhousetraces:
     datasource: tcp://clickhouse:9000/signoz_traces
     low_cardinal_exception_grouping: ${env:LOW_CARDINAL_EXCEPTION_GROUPING}
-    use_new_schema: true
   clickhousemetricswrite:
     endpoint: tcp://clickhouse:9000/signoz_metrics
     resource_to_telemetry_conversion:

deploy/docker/clickhouse-setup/docker-compose-local.yaml

@@ -25,8 +25,7 @@ services:
     command:
       [
         "-config=/root/config/prometheus.yml",
-        "--use-logs-new-schema=true",
-        "--use-trace-new-schema=true"
+        "--use-logs-new-schema=true"
       ]
     ports:
       - "6060:6060"

deploy/docker/clickhouse-setup/docker-compose-minimal.yaml

@@ -167,8 +167,7 @@ services:
     command:
       [
         "-config=/root/config/prometheus.yml",
-        "--use-logs-new-schema=true",
-        "--use-trace-new-schema=true"
+        "--use-logs-new-schema=true"
       ]
     # ports:
     #   - "6060:6060"     # pprof port

deploy/docker/clickhouse-setup/docker-compose.testing.yaml

@@ -173,8 +173,7 @@ services:
       [
         "-config=/root/config/prometheus.yml",
         "-gateway-url=https://api.staging.signoz.cloud",
-        "--use-logs-new-schema=true",
-        "--use-trace-new-schema=true"
+        "--use-logs-new-schema=true"
       ]
     # ports:
     #   - "6060:6060"     # pprof port

deploy/docker/clickhouse-setup/otel-collector-config.yaml

@@ -119,7 +119,6 @@ exporters:
   clickhousetraces:
     datasource: tcp://clickhouse:9000/signoz_traces
     low_cardinal_exception_grouping: ${env:LOW_CARDINAL_EXCEPTION_GROUPING}
-    use_new_schema: true
   clickhousemetricswrite:
     endpoint: tcp://clickhouse:9000/signoz_metrics
     resource_to_telemetry_conversion:

pkg/query-service/app/integrations/messagingQueues/kafka/sql.go

@@ -9,25 +9,25 @@ func generateConsumerSQL(start, end int64, topic, partition, consumerGroup, queu
 	query := fmt.Sprintf(`
 WITH consumer_query AS (
     SELECT
-        resource_string_service$$name,
+        serviceName,
         quantile(0.99)(durationNano) / 1000000 AS p99,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count,
-        avg(CASE WHEN has(attributes_number, 'messaging.message.body.size') THEN attributes_number['messaging.message.body.size'] ELSE NULL END) AS avg_msg_size
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count,
+        avg(CASE WHEN has(numberTagMap, 'messaging.message.body.size') THEN numberTagMap['messaging.message.body.size'] ELSE NULL END) AS avg_msg_size
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 5
-        AND attribute_string_messaging$$system = '%s'
-        AND attributes_string['messaging.destination.name'] = '%s'
-        AND attributes_string['messaging.destination.partition.id'] = '%s'
-        AND attributes_string['messaging.kafka.consumer.group'] = '%s'
-    GROUP BY resource_string_service$$name
+        AND msgSystem = '%s'
+        AND stringTagMap['messaging.destination.name'] = '%s'
+        AND stringTagMap['messaging.destination.partition.id'] = '%s'
+        AND stringTagMap['messaging.kafka.consumer.group'] = '%s'
+    GROUP BY serviceName
 )
 
 SELECT
-    resource_string_service$$name AS service_name,
+    serviceName AS service_name,
     p99,
     COALESCE((error_count * 100.0) / total_requests, 0) AS error_rate,
     COALESCE(total_requests / %d, 0) AS throughput,
@@ -35,7 +35,7 @@ SELECT
 FROM
     consumer_query
 ORDER BY
-    resource_string_service$$name;
+    serviceName;
 `, start, end, queueType, topic, partition, consumerGroup, timeRange)
 	return query
 }
@@ -48,14 +48,14 @@ WITH partition_query AS (
     SELECT
         quantile(0.99)(durationNano) / 1000000 AS p99,
         count(*) AS total_requests,
-        attributes_string['messaging.destination.name'] AS topic,
-		attributes_string['messaging.destination.partition.id'] AS partition
-    FROM signoz_traces.distributed_signoz_index_v3
+        stringTagMap['messaging.destination.name'] AS topic,
+		stringTagMap['messaging.destination.partition.id'] AS partition
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 4
-        AND attribute_string_messaging$$system = '%s'
+        AND msgSystem = '%s'
     GROUP BY topic, partition
 )
 
@@ -78,25 +78,25 @@ func generateConsumerPartitionLatencySQL(start, end int64, topic, partition, que
 	query := fmt.Sprintf(`
 WITH consumer_pl AS (
     SELECT
-        attributes_string['messaging.kafka.consumer.group'] AS consumer_group,
-        resource_string_service$$name,
+        stringTagMap['messaging.kafka.consumer.group'] AS consumer_group,
+        serviceName,
         quantile(0.99)(durationNano) / 1000000 AS p99,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 5
-        AND attribute_string_messaging$$system = '%s'
-        AND attributes_string['messaging.destination.name'] = '%s'
-        AND attributes_string['messaging.destination.partition.id'] = '%s'
-    GROUP BY consumer_group, resource_string_service$$name
+        AND msgSystem = '%s'
+        AND stringTagMap['messaging.destination.name'] = '%s'
+        AND stringTagMap['messaging.destination.partition.id'] = '%s'
+    GROUP BY consumer_group, serviceName
 )
 
 SELECT
     consumer_group,
-    resource_string_service$$name AS service_name,
+    serviceName AS service_name,
     p99,
     COALESCE((error_count * 100.0) / total_requests, 0) AS error_rate,
     COALESCE(total_requests / %d, 0) AS throughput
@@ -115,23 +115,23 @@ func generateProducerPartitionThroughputSQL(start, end int64, queueType string)
 	query := fmt.Sprintf(`
 WITH producer_latency AS (
     SELECT
-		resource_string_service$$name,
+		serviceName,
         quantile(0.99)(durationNano) / 1000000 AS p99,
-		attributes_string['messaging.destination.name'] AS topic,
+		stringTagMap['messaging.destination.name'] AS topic,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 4
-        AND attribute_string_messaging$$system = '%s'
-    GROUP BY topic, resource_string_service$$name
+        AND msgSystem = '%s'
+    GROUP BY topic, serviceName
 )
 
 SELECT
 	topic,
-	resource_string_service$$name AS service_name,
+	serviceName AS service_name,
     p99,
     COALESCE((error_count * 100.0) / total_requests, 0) AS error_rate,
     COALESCE(total_requests / %d, 0) AS throughput
@@ -148,17 +148,17 @@ func generateProducerTopicLatencySQL(start, end int64, topic, service, queueType
 WITH consumer_latency AS (
     SELECT
         quantile(0.99)(durationNano) / 1000000 AS p99,
-		attributes_string['messaging.destination.partition.id'] AS partition,
+		stringTagMap['messaging.destination.partition.id'] AS partition,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 4
-		AND resource_string_service$$name = '%s'
-        AND attribute_string_messaging$$system = '%s'
-		AND attributes_string['messaging.destination.name'] = '%s'
+		AND serviceName = '%s'
+        AND msgSystem = '%s'
+		AND stringTagMap['messaging.destination.name'] = '%s'
     GROUP BY partition
 )
 
@@ -179,24 +179,24 @@ func generateConsumerLatencySQL(start, end int64, queueType string) string {
 	query := fmt.Sprintf(`
 WITH consumer_latency AS (
     SELECT
-        resource_string_service$$name,
-        attributes_string['messaging.destination.name'] AS topic,
+        serviceName,
+        stringTagMap['messaging.destination.name'] AS topic,
         quantile(0.99)(durationNano) / 1000000 AS p99,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count,
-        SUM(attributes_number['messaging.message.body.size']) AS total_bytes
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count,
+        SUM(numberTagMap['messaging.message.body.size']) AS total_bytes
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 5
-        AND attribute_string_messaging$$system = '%s'
-    GROUP BY topic, resource_string_service$$name
+        AND msgSystem = '%s'
+    GROUP BY topic, serviceName
 )
 
 SELECT
     topic,
-    resource_string_service$$name AS service_name,
+    serviceName AS service_name,
     p99,
     COALESCE((error_count * 100.0) / total_requests, 0) AS error_rate,
     COALESCE(total_requests / %d, 0) AS ingestion_rate,
@@ -216,17 +216,17 @@ func generateConsumerServiceLatencySQL(start, end int64, topic, service, queueTy
 WITH consumer_latency AS (
     SELECT
         quantile(0.99)(durationNano) / 1000000 AS p99,
-		attributes_string['messaging.destination.partition.id'] AS partition,
+		stringTagMap['messaging.destination.partition.id'] AS partition,
         COUNT(*) AS total_requests,
-        sumIf(1, status_code = 2) AS error_count
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 5
-		AND resource_string_service$$name = '%s'
-        AND attribute_string_messaging$$system = '%s'
-		AND attributes_string['messaging.destination.name'] = '%s'
+		AND serviceName = '%s'
+        AND msgSystem = '%s'
+		AND stringTagMap['messaging.destination.name'] = '%s'
     GROUP BY partition
 )
 
@@ -246,26 +246,26 @@ func generateProducerConsumerEvalSQL(start, end int64, queueType string, evalTim
 	query := fmt.Sprintf(`
 WITH trace_data AS (
     SELECT
-        p.resource_string_service$$name AS producer_service,
-        c.resource_string_service$$name AS consumer_service,
-        p.trace_id,
+        p.serviceName AS producer_service,
+        c.serviceName AS consumer_service,
+        p.traceID,
         p.timestamp AS producer_timestamp,
         c.timestamp AS consumer_timestamp,
         p.durationNano AS durationNano,
         (toUnixTimestamp64Nano(c.timestamp) - toUnixTimestamp64Nano(p.timestamp)) + p.durationNano AS time_difference
     FROM
-        signoz_traces.distributed_signoz_index_v3 p
+        signoz_traces.distributed_signoz_index_v2 p
     INNER JOIN
-        signoz_traces.distributed_signoz_index_v3 c
-            ON p.trace_id = c.trace_id
-            AND c.parent_span_id = p.span_id
+        signoz_traces.distributed_signoz_index_v2 c
+            ON p.traceID = c.traceID
+            AND c.parentSpanID = p.spanID
     WHERE
         p.kind = 4
         AND c.kind = 5
         AND toUnixTimestamp64Nano(p.timestamp) BETWEEN '%d' AND '%d'
         AND toUnixTimestamp64Nano(c.timestamp) BETWEEN '%d' AND '%d'
-        AND c.attribute_string_messaging$$system = '%s'
-        AND p.attribute_string_messaging$$system = '%s'
+        AND c.msgSystem = '%s'
+        AND p.msgSystem = '%s'
 )
 
 SELECT
@@ -278,7 +278,7 @@ SELECT
         arrayMap(x -> x.1,
             arraySort(
                 x -> -x.2,
-                groupArrayIf((trace_id, time_difference), time_difference > '%d')
+                groupArrayIf((traceID, time_difference), time_difference > '%d')
             )
         ),
         1, 10
@@ -296,30 +296,30 @@ func generateProducerSQL(start, end int64, topic, partition, queueType string) s
 	query := fmt.Sprintf(`
 WITH producer_query AS (
     SELECT
-        resource_string_service$$name,
+        serviceName,
         quantile(0.99)(durationNano) / 1000000 AS p99,
         count(*) AS total_count,
-        sumIf(1, status_code = 2) AS error_count
-    FROM signoz_traces.distributed_signoz_index_v3
+        sumIf(1, statusCode = 2) AS error_count
+    FROM signoz_traces.distributed_signoz_index_v2
     WHERE
         timestamp >= '%d'
         AND timestamp <= '%d'
         AND kind = 4
-        AND attribute_string_messaging$$system = '%s'
-        AND attributes_string['messaging.destination.name'] = '%s'
-        AND attributes_string['messaging.destination.partition.id'] = '%s'
-    GROUP BY resource_string_service$$name
+        AND msgSystem = '%s'
+        AND stringTagMap['messaging.destination.name'] = '%s'
+        AND stringTagMap['messaging.destination.partition.id'] = '%s'
+    GROUP BY serviceName
 )
 
 SELECT
-    resource_string_service$$name AS service_name,
+    serviceName AS service_name,
     p99,
     COALESCE((error_count * 100.0) / total_count, 0) AS error_percentage,
     COALESCE(total_count / %d, 0) AS throughput
 FROM
     producer_query
 ORDER BY
-    resource_string_service$$name;
+    serviceName;
 `, start, end, queueType, topic, partition, timeRange)
 	return query
 }
@@ -328,18 +328,18 @@ func generateNetworkLatencyThroughputSQL(start, end int64, consumerGroup, partit
 	timeRange := (end - start) / 1000000000
 	query := fmt.Sprintf(`
 SELECT
-    attributes_string['messaging.client_id'] AS client_id,
-	attributes_string['service.instance.id'] AS service_instance_id,
-    resource_string_service$$name AS service_name,
+    stringTagMap['messaging.client_id'] AS client_id,
+	stringTagMap['service.instance.id'] AS service_instance_id,
+    serviceName AS service_name,
     count(*) / %d AS throughput
-FROM signoz_traces.distributed_signoz_index_v3
+FROM signoz_traces.distributed_signoz_index_v2
 WHERE
     timestamp >= '%d'
     AND timestamp <= '%d'
     AND kind = 5
-    AND attribute_string_messaging$$system = '%s' 
-    AND attributes_string['messaging.kafka.consumer.group'] = '%s'
-    AND attributes_string['messaging.destination.partition.id'] = '%s'
+    AND msgSystem = '%s' 
+    AND stringTagMap['messaging.kafka.consumer.group'] = '%s'
+    AND stringTagMap['messaging.destination.partition.id'] = '%s'
 GROUP BY service_name, client_id, service_instance_id
 ORDER BY throughput DESC
 `, timeRange, start, end, queueType, consumerGroup, partitionID)
@@ -350,12 +350,12 @@ func onboardProducersSQL(start, end int64, queueType string) string {
 	query := fmt.Sprintf(`
 SELECT 
     COUNT(*) = 0 AS entries,
-    COUNT(IF(attribute_string_messaging$$system = '%s', 1, NULL)) = 0 AS queue,
+    COUNT(IF(msgSystem = '%s', 1, NULL)) = 0 AS queue,
     COUNT(IF(kind = 4, 1, NULL)) = 0 AS kind,
-    COUNT(IF(has(attributes_string, 'messaging.destination.name'), 1, NULL)) = 0 AS destination,
-    COUNT(IF(has(attributes_string, 'messaging.destination.partition.id'), 1, NULL)) = 0 AS partition
+    COUNT(IF(has(stringTagMap, 'messaging.destination.name'), 1, NULL)) = 0 AS destination,
+    COUNT(IF(has(stringTagMap, 'messaging.destination.partition.id'), 1, NULL)) = 0 AS partition
 FROM 
-    signoz_traces.distributed_signoz_index_v3
+    signoz_traces.distributed_signoz_index_v2
 WHERE 
     timestamp >= '%d'
     AND timestamp <= '%d';`, queueType, start, end)
@@ -366,16 +366,16 @@ func onboardConsumerSQL(start, end int64, queueType string) string {
 	query := fmt.Sprintf(`
 SELECT  
     COUNT(*) = 0 AS entries,
-    COUNT(IF(attribute_string_messaging$$system = '%s', 1, NULL)) = 0 AS queue,
+    COUNT(IF(msgSystem = '%s', 1, NULL)) = 0 AS queue,
     COUNT(IF(kind = 5, 1, NULL)) = 0 AS kind,
-    COUNT(resource_string_service$$name) = 0 AS svc,
-    COUNT(IF(has(attributes_string, 'messaging.destination.name'), 1, NULL)) = 0 AS destination,
-    COUNT(IF(has(attributes_string, 'messaging.destination.partition.id'), 1, NULL)) = 0 AS partition,
-    COUNT(IF(has(attributes_string, 'messaging.kafka.consumer.group'), 1, NULL)) = 0 AS cgroup,
-    COUNT(IF(has(attributes_number, 'messaging.message.body.size'), 1, NULL)) = 0 AS bodysize,
-    COUNT(IF(has(attributes_string, 'messaging.client_id'), 1, NULL)) = 0 AS clientid,
-    COUNT(IF(has(attributes_string, 'service.instance.id'), 1, NULL)) = 0 AS instanceid
-FROM signoz_traces.distributed_signoz_index_v3
+    COUNT(serviceName) = 0 AS svc,
+    COUNT(IF(has(stringTagMap, 'messaging.destination.name'), 1, NULL)) = 0 AS destination,
+    COUNT(IF(has(stringTagMap, 'messaging.destination.partition.id'), 1, NULL)) = 0 AS partition,
+    COUNT(IF(has(stringTagMap, 'messaging.kafka.consumer.group'), 1, NULL)) = 0 AS cgroup,
+    COUNT(IF(has(numberTagMap, 'messaging.message.body.size'), 1, NULL)) = 0 AS bodysize,
+    COUNT(IF(has(stringTagMap, 'messaging.client_id'), 1, NULL)) = 0 AS clientid,
+    COUNT(IF(has(stringTagMap, 'service.instance.id'), 1, NULL)) = 0 AS instanceid
+FROM signoz_traces.distributed_signoz_index_v2
 WHERE 
     timestamp >= '%d'
     AND timestamp <= '%d';`, queueType, start, end)

pkg/query-service/app/queryBuilder/query_builder.go

@@ -41,7 +41,7 @@ var SupportedFunctions = []string{
 
 var EvalFuncs = map[string]govaluate.ExpressionFunction{}
 
-type prepareTracesQueryFunc func(start, end int64, panelType v3.PanelType, bq *v3.BuilderQuery, options v3.QBOptions) (string, error)
+type prepareTracesQueryFunc func(start, end int64, panelType v3.PanelType, bq *v3.BuilderQuery, options v3.QBOptions, sortBy string) (string, error)
 type prepareLogsQueryFunc func(start, end int64, queryType v3.QueryType, panelType v3.PanelType, bq *v3.BuilderQuery, options v3.QBOptions) (string, error)
 type prepareMetricQueryFunc func(start, end int64, queryType v3.QueryType, panelType v3.PanelType, bq *v3.BuilderQuery, options metricsV3.Options) (string, error)
 
@@ -196,12 +196,12 @@ func (qb *QueryBuilder) PrepareQueries(params *v3.QueryRangeParamsV3) (map[strin
 					// for ts query with group by and limit form two queries
 					if compositeQuery.PanelType == v3.PanelTypeGraph && query.Limit > 0 && len(query.GroupBy) > 0 {
 						limitQuery, err := qb.options.BuildTraceQuery(start, end, compositeQuery.PanelType, query,
-							v3.QBOptions{GraphLimitQtype: constants.FirstQueryGraphLimit, PreferRPM: PreferRPMFeatureEnabled})
+							v3.QBOptions{GraphLimitQtype: constants.FirstQueryGraphLimit, PreferRPM: PreferRPMFeatureEnabled}, params.SortBy)
 						if err != nil {
 							return nil, err
 						}
 						placeholderQuery, err := qb.options.BuildTraceQuery(start, end, compositeQuery.PanelType,
-							query, v3.QBOptions{GraphLimitQtype: constants.SecondQueryGraphLimit, PreferRPM: PreferRPMFeatureEnabled})
+							query, v3.QBOptions{GraphLimitQtype: constants.SecondQueryGraphLimit, PreferRPM: PreferRPMFeatureEnabled}, params.SortBy)
 						if err != nil {
 							return nil, err
 						}
@@ -209,7 +209,7 @@ func (qb *QueryBuilder) PrepareQueries(params *v3.QueryRangeParamsV3) (map[strin
 						queries[queryName] = query
 					} else {
 						queryString, err := qb.options.BuildTraceQuery(start, end, compositeQuery.PanelType,
-							query, v3.QBOptions{PreferRPM: PreferRPMFeatureEnabled, GraphLimitQtype: ""})
+							query, v3.QBOptions{PreferRPM: PreferRPMFeatureEnabled, GraphLimitQtype: ""}, params.SortBy)
 						if err != nil {
 							return nil, err
 						}

pkg/query-service/app/traces/v3/query_builder.go

@@ -499,7 +499,7 @@ func AddOffsetToQuery(query string, offset uint64) string {
 // PrepareTracesQuery returns the query string for traces
 // start and end are in epoch millisecond
 // step is in seconds
-func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.BuilderQuery, options v3.QBOptions) (string, error) {
+func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.BuilderQuery, options v3.QBOptions, sortBy string) (string, error) {
 	// adjust the start and end time to the step interval
 	if panelType == v3.PanelTypeGraph {
 		// adjust the start and end time to the step interval for graph panel types

pkg/query-service/app/traces/v4/query_builder.go

@@ -200,7 +200,7 @@ func orderByAttributeKeyTags(panelType v3.PanelType, items []v3.OrderBy, tags []
 	return str
 }
 
-func buildTracesQuery(start, end, step int64, mq *v3.BuilderQuery, panelType v3.PanelType, options v3.QBOptions) (string, error) {
+func buildTracesQuery(start, end, step int64, mq *v3.BuilderQuery, panelType v3.PanelType, options v3.QBOptions, sortBy string) (string, error) {
 	tracesStart := utils.GetEpochNanoSecs(start)
 	tracesEnd := utils.GetEpochNanoSecs(end)
 
@@ -249,12 +249,21 @@ func buildTracesQuery(start, end, step int64, mq *v3.BuilderQuery, panelType v3.
 	if mq.AggregateOperator == v3.AggregateOperatorNoOp {
 		var query string
 		if panelType == v3.PanelTypeTrace {
-			withSubQuery := fmt.Sprintf(constants.TracesExplorerViewSQLSelectWithSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3_LOCAL_TABLENAME, timeFilter, filterSubQuery)
-			withSubQuery = tracesV3.AddLimitToQuery(withSubQuery, mq.Limit)
-			if mq.Offset != 0 {
-				withSubQuery = tracesV3.AddOffsetToQuery(withSubQuery, mq.Offset)
+			if sortBy == constants.TraceSortBySpanCount {
+				spanCountSubQuery := fmt.Sprintf(constants.TraceExplorerSpanCountSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_TRACE_SPAN_COUNT_MV, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3_LOCAL_TABLENAME, timeFilter, filterSubQuery)
+				query = tracesV3.AddLimitToQuery(spanCountSubQuery, mq.Limit)
+				if mq.Offset != 0 {
+					query = tracesV3.AddOffsetToQuery(query, mq.Offset)
+				}
+			} else {
+				withSubQuery := fmt.Sprintf(constants.TracesExplorerViewSQLSelectWithSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3_LOCAL_TABLENAME, timeFilter, filterSubQuery)
+				withSubQuery = tracesV3.AddLimitToQuery(withSubQuery, mq.Limit)
+				if mq.Offset != 0 {
+					withSubQuery = tracesV3.AddOffsetToQuery(withSubQuery, mq.Offset)
+				}
+				query = fmt.Sprintf(constants.TracesExplorerViewSQLSelectBeforeSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3) + withSubQuery + ") " + fmt.Sprintf(constants.TracesExplorerViewSQLSelectAfterSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3, timeFilter)
+
 			}
-			query = fmt.Sprintf(constants.TracesExplorerViewSQLSelectBeforeSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3) + withSubQuery + ") " + fmt.Sprintf(constants.TracesExplorerViewSQLSelectAfterSubQuery, constants.SIGNOZ_TRACE_DBNAME, constants.SIGNOZ_SPAN_INDEX_V3, timeFilter)
 			// adding this to avoid the distributed product mode error which doesn't allow global in
 			query += " settings distributed_product_mode='allow', max_memory_usage=10000000000"
 		} else if panelType == v3.PanelTypeList {
@@ -375,7 +384,7 @@ func buildTracesQuery(start, end, step int64, mq *v3.BuilderQuery, panelType v3.
 // PrepareTracesQuery returns the query string for traces
 // start and end are in epoch millisecond
 // step is in seconds
-func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.BuilderQuery, options v3.QBOptions) (string, error) {
+func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.BuilderQuery, options v3.QBOptions, sortBy string) (string, error) {
 	// adjust the start and end time to the step interval
 	if panelType == v3.PanelTypeGraph {
 		// adjust the start and end time to the step interval for graph panel types
@@ -384,7 +393,7 @@ func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.Builder
 	}
 	if options.GraphLimitQtype == constants.FirstQueryGraphLimit {
 		// give me just the group by names
-		query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options)
+		query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options, sortBy)
 		if err != nil {
 			return "", err
 		}
@@ -392,14 +401,14 @@ func PrepareTracesQuery(start, end int64, panelType v3.PanelType, mq *v3.Builder
 
 		return query, nil
 	} else if options.GraphLimitQtype == constants.SecondQueryGraphLimit {
-		query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options)
+		query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options, sortBy)
 		if err != nil {
 			return "", err
 		}
 		return query, nil
 	}
 
-	query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options)
+	query, err := buildTracesQuery(start, end, mq.StepInterval, mq, panelType, options, sortBy)
 	if err != nil {
 		return "", err
 	}

pkg/query-service/constants/constants.go

@@ -53,6 +53,8 @@ const LogsTTL = "logs"
 const DurationSort = "DurationSort"
 const TimestampSort = "TimestampSort"
 const PreferRPM = "PreferRPM"
+const TraceSortBySpanCount = "span_count"
+const TraceSortByTraceDuration = "trace_duration"
 
 func GetAlertManagerApiPrefix() string {
 	if os.Getenv("ALERTMANAGER_API_PREFIX") != "" {
@@ -90,7 +92,7 @@ func EnableHostsInfraMonitoring() bool {
 	return GetOrDefaultEnv("ENABLE_INFRA_METRICS", "true") == "true"
 }
 
-var KafkaSpanEval = GetOrDefaultEnv("KAFKA_SPAN_EVAL", "true")
+var KafkaSpanEval = GetOrDefaultEnv("KAFKA_SPAN_EVAL", "false")
 
 func IsDurationSortFeatureEnabled() bool {
 	isDurationSortFeatureEnabledStr := DurationSortFeature
@@ -248,6 +250,7 @@ const (
 	SIGNOZ_TIMESERIES_v4_1DAY_LOCAL_TABLENAME  = "time_series_v4_1day"
 	SIGNOZ_TIMESERIES_v4_1WEEK_LOCAL_TABLENAME = "time_series_v4_1week"
 	SIGNOZ_TIMESERIES_v4_1DAY_TABLENAME        = "distributed_time_series_v4_1day"
+	SIGNOZ_TRACE_SPAN_COUNT_MV                 = "trace_summary_mv"
 )
 
 var TimeoutExcludedRoutes = map[string]bool{
@@ -353,6 +356,8 @@ const (
 	TracesExplorerViewSQLSelectQuery = "SELECT subQuery.serviceName, subQuery.name, count() AS " +
 		"span_count, subQuery.durationNano, traceID FROM %s.%s GLOBAL INNER JOIN subQuery ON %s.traceID = subQuery.traceID GROUP " +
 		"BY traceID, subQuery.durationNano, subQuery.name, subQuery.serviceName ORDER BY subQuery.durationNano desc;"
+	TraceExplorerSpanCountSubQuery    = "SELECT t.trace_id, t.num_spans FROM %s.%s as t GLOBAL INNER JOIN (SELECT * FROM (SELECT trace_id FROM %s.%s WHERE %s %s LIMIT 1 BY trace_id) AS inner_filtered ) AS filtered_traces ON t.trace_id = filtered_traces.trace_id ORDER BY t.num_spans DESC "
+	TraceExplorerEnrichSpanCountQuery = "SELECT name, serviceName, duration_nano FROM %s.%s WHERE trace_id = %s"
 )
 
 // ReservedColumnTargetAliases identifies result value from a user

pkg/query-service/model/v3/v3.go

@@ -379,6 +379,7 @@ type QueryRangeParamsV3 struct {
 	NoCache        bool                   `json:"noCache"`
 	Version        string                 `json:"-"`
 	FormatForWeb   bool                   `json:"formatForWeb,omitempty"`
+	SortBy         string                 `json:"sortBy,omitempty"`
 }
 
 func (q *QueryRangeParamsV3) Clone() *QueryRangeParamsV3 {