SigNoz/signoz · error

error while fetching trace schema: %s

Error message

error while fetching trace schema: %s

What it means

After successfully listing tag keys, the reader runs SHOW CREATE TABLE <tracesdb>.<traceLocalTableName> via db.Select into []model.ShowCreateTableStatement and that failed. The schema string is needed to compute the IsColumn flag for each attribute key, so the whole request fails.

Source

Thrown at pkg/query-service/app/clickhouseReader/reader.go:4241

		return nil, fmt.Errorf("unsupported aggregate operator")
	}
	query = fmt.Sprintf("SELECT DISTINCT(tag_key), tag_type, tag_data_type FROM %s.%s WHERE %s and tag_type != 'spanfield'", r.TraceDB, r.spanAttributeTableV2, where)
	if req.Limit != 0 {
		query = query + fmt.Sprintf(" LIMIT %d;", req.Limit)
	}
	rows, err = r.db.Query(ctx, query, fmt.Sprintf("%%%s%%", req.SearchText))

	if err != nil {
		r.logger.Error("Error while executing query", errorsV2.Attr(err))
		return nil, fmt.Errorf("error while executing query: %s", err.Error())
	}
	defer rows.Close()

	statements := []model.ShowCreateTableStatement{}
	query = fmt.Sprintf("SHOW CREATE TABLE %s.%s", r.TraceDB, r.traceLocalTableName)
	err = r.db.Select(ctx, &statements, query)
	if err != nil {
		return nil, fmt.Errorf("error while fetching trace schema: %s", err.Error())
	}

	var tagKey string
	var dataType string
	var tagType string
	for rows.Next() {
		if err := rows.Scan(&tagKey, &tagType, &dataType); err != nil {
			return nil, fmt.Errorf("error while scanning rows: %s", err.Error())
		}
		key := v3.AttributeKey{
			Key:      tagKey,
			DataType: v3.AttributeKeyDataType(dataType),
			Type:     v3.AttributeKeyType(tagType),
			IsColumn: isColumn(statements[0].Statement, tagType, tagKey, dataType),
		}

		if _, ok := constants.DeprecatedStaticFieldsTraces[tagKey]; !ok {
			response.AttributeKeys = append(response.AttributeKeys, key)

View on GitHub (pinned to 5069bf80b0)

Solutions

  1. Run SHOW CREATE TABLE <tracesdb>.<traceLocalTableName> manually as the query-service ClickHouse user to reproduce the exact error
  2. Check the env/config value for the local trace table name and align it with the actual local table (SHOW TABLES FROM <tracesdb>)
  3. Grant the ClickHouse user SHOW/SELECT privileges if the error is permission related
  4. Update SigNoz so table naming matches the current schema migrations
Defensive patterns

Strategy: try-catch

Validate before calling

// startup check: verify the local trace table is visible
rows, err := ch.Query(ctx, "SHOW TABLES FROM "+traceDB+" LIKE '"+localTable+"'")
if err != nil || !rows.Next() {
  log.Fatal("local trace table not found; check CH_LOCAL_EVENT_TABLE config")
}

Try / catch

if err != nil {
  if strings.Contains(err.Error(), "error while fetching trace schema") {
    // schema lookup failed: check table existence/permissions, return 503 not 400
  }
}

Prevention

When it happens

Trigger: Calling GetAggregateAttribute on a distributed SigNoz setup where the local table (e.g. distributed_signoz_trace_spans / signoz_trace_spans_local) named by r.traceLocalTableName does not exist on the node being queried; insufficient permissions for SHOW CREATE TABLE; ClickHouse restarted mid-request dropping the connection.

Common situations: Config variable CH_LOCAL_EVENT_TABLE / traceLocalTableName set to a wrong name; a migration renamed the local table; user account lacks SHOW grant; older cluster where only the Distributed table exists with a different local name.

Related errors

AI-assisted analysis of SigNoz/signoz@5069bf80b0 (2026-08-28). Data as JSON: /api/errors/24f54d0e70ca486c. Report an issue: GitHub.