jaegertracing/jaeger · error

error reading operation_names from storage: %w

Error message

error reading operation_names from storage: %w

What it means

getOperationsV1 scans all rows from the legacy operation_names table via gocql's iterator; if iter.Close() returns an error (query failed mid-scan, timeout, node down), the raw driver error is wrapped with this message to give context about which read failed.

Source

Thrown at internal/storage/v1/cassandra/spanstore/operation_names.go:174

	err := query.Exec()
	return err == nil
}

func getOperationsV1(
	s *OperationNamesStorage,
	query tracestore.OperationQueryParams,
) ([]tracestore.Operation, error) {
	iter := s.session.Query(s.table.queryStmt, query.ServiceName).Iter()

	var operation string
	var operations []tracestore.Operation
	for iter.Scan(&operation) {
		operations = append(operations, tracestore.Operation{
			Name: operation,
		})
	}
	if err := iter.Close(); err != nil {
		err = fmt.Errorf("error reading operation_names from storage: %w", err)
		return nil, err
	}

	return operations, nil
}

func getOperationsV2(
	s *OperationNamesStorage,
	query tracestore.OperationQueryParams,
) ([]tracestore.Operation, error) {
	var casQuery cassandra.Query
	if query.SpanKind == "" {
		// Get operations for all spanKind
		casQuery = s.session.Query(s.table.queryStmt, query.ServiceName)
	} else {
		// Get operations for given spanKind
		casQuery = s.session.Query(s.table.queryByKindStmt, query.ServiceName, query.SpanKind)
	}

View on GitHub (pinned to 806f444784)

Solutions

1. Check Cassandra logs / system.traces for the underlying read failure (timeouts, unavailable) and address cluster health.
2. Upgrade the schema to operation_names_v2 (run latest schema script), which is bucketed and less timeout-prone.
3. Retry the GetServices/GetOperations call once the cluster recovers; it is a read-only operation.
4. Tune Cassandra client timeouts/retry policy in the Jaeger storage configuration if timeouts are frequent.

ops, err := reader.GetOperations(ctx, tracestore.OperationQuery{ServiceName: svc})
if err != nil {
    if strings.Contains(err.Error(), "error reading operation_names") {
        // transient Cassandra read failure: backoff and retry
    }
    return err
}

When it happens

Trigger: Calling GetOperations (SpanReader) against a keyspace on schema version 1 (operation_names table) while the underlying Cassandra query fails — coordinator timeout, node unavailability, tombstone/gc_grace issues during a full-scan query.

Common situations: Cassandra cluster under load causing read timeouts on the unbucketed v1 table; network blips between Jaeger and Cassandra; large per-service operation lists hitting read timeouts.

Related errors

• error reading %s from storage: %w
• error reading traces from storage: %w
• error reading service_names from storage: %w
• invalid version
• failed to Exec query '%s': %w

AI-assisted analysis of jaegertracing/jaeger@806f444784 (2026-09-01). Data as JSON: /api/errors/989dc194378433db. Report an issue: GitHub.