SigNoz/signoz · error · model.ApiError

couldn't query attrib values for suggestions: %w

Error message

couldn't query attrib values for suggestions: %w

What it means

InternalError from getValuesForLogAttributes when the UNION DISTINCT tag-value query against ClickHouse fails while building log filter suggestions. The query is built per requested attribute and executed with max_threads=2; any DB-level failure (timeout, missing column, overload) triggers it.

Source

Thrown at pkg/query-service/app/clickhouseReader/filter_suggestions.go:188

			select tag_key, string_value, number_value
			from %s.%s
			where tag_key = $%d and (
				string_value != '' or number_value is not null
			) and tag_type != 'logfield'
			limit %d
		)`, r.logsDB, r.logsTagAttributeTableV2, idx+1, limit))

		tagKeyQueryArgs = append(tagKeyQueryArgs, attrib.Key)
	}

	query := fmt.Sprintf(`select * from (
		%s
	) settings max_threads=2`, strings.Join(tagQueries, " UNION DISTINCT "))

	rows, err := r.db.Query(ctx, query, tagKeyQueryArgs...)
	if err != nil {
		r.logger.ErrorContext(ctx, "couldn't query attrib values for suggestions", errorsV2.Attr(err))
		return nil, model.InternalError(fmt.Errorf(
			"couldn't query attrib values for suggestions: %w", err,
		))
	}
	defer rows.Close()

	result := make([][]any, len(attributes))

	// Helper for getting hold of the result slice to append to for each scanned row
	resultIdxForAttrib := func(key string, dataType v3.AttributeKeyDataType) int {
		return slices.IndexFunc(attributes, func(attrib v3.AttributeKey) bool {
			return attrib.Key == key && attrib.DataType == dataType
		})
	}

	// Scan rows and append to result
	for rows.Next() {
		var tagKey string
		var stringValue string

View on GitHub (pinned to 5069bf80b0)

Solutions

1. Check server logs for the logged ClickHouse error (it is logged just before returning)
2. Validate the requested attribute keys actually exist as columns in the log tables
3. Reduce the number/range of attributes requested or add time bounds
4. Scale/optimize ClickHouse (marks, memory limits) or retry when load drops

// Validate requested attributes exist before asking for values
for _, k := range req.AttributeKeys {
    if !attributeKeyExists(ctx, ch, k) { return fmt.Errorf("unknown attribute %s", k) }
}

func isChQueryErr(err error) bool {
    var apiErr *model.ApiError
    return errors.As(err, &apiErr) && apiErr.Typ == model.ErrorInternal && strings.Contains(err.Error(), "attrib values")
}

vals, err := reader.GetQBFilterSuggestionsForLogs(ctx, req)
if err != nil && isChQueryErr(err) {
    // transient CH failures are common; retry once with smaller attribute set
    req.AttributeKeys = req.AttributeKeys[:1]
    vals, err = reader.GetQBFilterSuggestionsForLogs(ctx, req)
}

When it happens

Trigger: GetQBFilterSuggestionsForLogs with one or more string/number attributes whose UNION-ed SELECT against the tag value tables fails: unknown column (renamed/missing attribute key), ClickHouse timeout, or packet/protocol errors under load.

Common situations: Attribute keys that no longer exist in the schema, mixed schema versions after upgrade, ClickHouse under memory pressure, very high-cardinality attributes making the UNION DISTINCT query slow/aborted.

Related errors

• couldn't get attribute keys: %w
• removing a selected field is not allowed, please reach out t
• couldn't scan attrib value rows: %w
• error in processing sql query
• error in processing sql query %w

AI-assisted analysis of SigNoz/signoz@5069bf80b0 (2026-08-28). Data as JSON: /api/errors/1463853c7d1691cb. Report an issue: GitHub.