fix: adding migration for fixing wrong cloud integration unique index

pkg/signoz/provider.go

@@ -172,6 +172,7 @@ func NewSQLMigrationProviderFactories(
 		sqlmigration.NewMigrateRulesV4ToV5Factory(sqlstore, telemetryStore),
 		sqlmigration.NewAddStatusUserFactory(sqlstore, sqlschema),
 		sqlmigration.NewDeprecateUserInviteFactory(sqlstore, sqlschema),
+		sqlmigration.NewFixCloudIntegrationUniqueIndexFactory(sqlstore, sqlschema),
 	)
 }
 

pkg/sqlmigration/069_fix_cloud_integration_index.go

@@ -0,0 +1,252 @@
+package sqlmigration
+
+import (
+	"context"
+	"database/sql"
+	"encoding/json"
+	"errors"
+	"time"
+
+	"github.com/SigNoz/signoz/pkg/factory"
+	"github.com/SigNoz/signoz/pkg/sqlschema"
+	"github.com/SigNoz/signoz/pkg/sqlstore"
+	"github.com/uptrace/bun"
+	"github.com/uptrace/bun/migrate"
+)
+
+type fixCloudIntegrationUniqueIndex struct {
+	sqlstore  sqlstore.SQLStore
+	sqlschema sqlschema.SQLSchema
+}
+
+func NewFixCloudIntegrationUniqueIndexFactory(sqlstore sqlstore.SQLStore, sqlschema sqlschema.SQLSchema) factory.ProviderFactory[SQLMigration, Config] {
+	return factory.NewProviderFactory(
+		factory.MustNewName("fix_cloud_integration_index"),
+		func(ctx context.Context, ps factory.ProviderSettings, c Config) (SQLMigration, error) {
+			return &fixCloudIntegrationUniqueIndex{
+				sqlstore:  sqlstore,
+				sqlschema: sqlschema,
+			}, nil
+		},
+	)
+}
+
+func (migration *fixCloudIntegrationUniqueIndex) Register(migrations *migrate.Migrations) error {
+	if err := migrations.Register(migration.Up, migration.Down); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+type cloudIntegrationRow struct {
+	bun.BaseModel `bun:"table:cloud_integration"`
+
+	ID        string    `bun:"id"`
+	AccountID string    `bun:"account_id"`
+	Provider  string    `bun:"provider"`
+	OrgID     string    `bun:"org_id"`
+	Config    string    `bun:"config"`
+	UpdatedAt time.Time `bun:"updated_at"`
+}
+
+type cloudIntegrationAccountConfig struct {
+	Regions []string `json:"regions"`
+}
+
+// duplicateGroup holds the keeper (first element) and losers (rest) for a duplicate (account_id, provider, org_id) group.
+type duplicateGroup struct {
+	keeper *cloudIntegrationRow
+	losers []*cloudIntegrationRow
+}
+
+func (migration *fixCloudIntegrationUniqueIndex) Up(ctx context.Context, db *bun.DB) error {
+	tx, err := db.BeginTx(ctx, nil)
+	if err != nil {
+		return err
+	}
+	defer func() {
+		_ = tx.Rollback()
+	}()
+
+	// Step 1: Drop the wrong index on (id, provider, org_id)
+	dropSqls := migration.sqlschema.Operator().DropIndex(
+		(&sqlschema.UniqueIndex{
+			TableName:   "cloud_integration",
+			ColumnNames: []sqlschema.ColumnName{"id", "provider", "org_id"},
+		}).Named("unique_cloud_integration"),
+	)
+	for _, sql := range dropSqls {
+		if _, err := tx.ExecContext(ctx, string(sql)); err != nil {
+			return err
+		}
+	}
+
+	// Step 2: Fetch all active rows with non-null account_id, ordered for grouping
+	var activeRows []*cloudIntegrationRow
+	err = tx.NewSelect().
+		Model(&activeRows).
+		Where("removed_at IS NULL").
+		Where("account_id IS NOT NULL").
+		OrderExpr("account_id, provider, org_id, updated_at DESC").
+		Scan(ctx)
+	if err != nil && !errors.Is(err, sql.ErrNoRows) {
+		return err
+	}
+
+	// Group by (account_id, provider, org_id)
+	groups := groupCloudIntegrationRows(activeRows)
+
+	now := time.Now()
+	var loserIDs []string
+
+	for _, group := range groups {
+		if len(group.losers) == 0 {
+			continue
+		}
+
+		// Step 3: Merge config from losers into keeper
+		if err = mergeCloudIntegrationConfigs(ctx, tx, group); err != nil {
+			return err
+		}
+
+		// Step 4: Reassign orphaned cloud_integration_service rows
+		for _, loser := range group.losers {
+			// Delete services from loser that would conflict with keeper's services (same type)
+			_, err = tx.NewDelete().
+				TableExpr("cloud_integration_service").
+				Where("cloud_integration_id = ?", loser.ID).
+				Where("type IN (?)",
+					tx.NewSelect().
+						TableExpr("cloud_integration_service").
+						Column("type").
+						Where("cloud_integration_id = ?", group.keeper.ID),
+				).
+				Exec(ctx)
+			if err != nil {
+				return err
+			}
+
+			// Reassign remaining services to the keeper
+			_, err = tx.ExecContext(ctx,
+				"UPDATE cloud_integration_service SET cloud_integration_id = ? WHERE cloud_integration_id = ?",
+				group.keeper.ID, loser.ID,
+			)
+			if err != nil {
+				return err
+			}
+
+			loserIDs = append(loserIDs, loser.ID)
+		}
+	}
+
+	// Step 5: Soft-delete all loser rows
+	if len(loserIDs) > 0 {
+		_, err = tx.NewUpdate().
+			TableExpr("cloud_integration").
+			Set("removed_at = ?", now).
+			Set("updated_at = ?", now).
+			Where("id IN (?)", bun.In(loserIDs)).
+			Exec(ctx)
+		if err != nil {
+			return err
+		}
+	}
+
+	// Step 6: Create correct partial unique index on (account_id, provider, org_id) WHERE removed_at IS NULL
+	createSqls := migration.sqlschema.Operator().CreateIndex(
+		&sqlschema.PartialUniqueIndex{
+			TableName:   "cloud_integration",
+			ColumnNames: []sqlschema.ColumnName{"account_id", "provider", "org_id"},
+			Where:       "removed_at IS NULL",
+		},
+	)
+	for _, sql := range createSqls {
+		if _, err = tx.ExecContext(ctx, string(sql)); err != nil {
+			return err
+		}
+	}
+
+	return tx.Commit()
+}
+
+func (migration *fixCloudIntegrationUniqueIndex) Down(ctx context.Context, db *bun.DB) error {
+	return nil
+}
+
+// groupCloudIntegrationRows groups rows by (account_id, provider, org_id).
+// Rows must be pre-sorted by account_id, provider, org_id, updated_at DESC
+// so the first row in each group is the keeper (most recently updated).
+func groupCloudIntegrationRows(rows []*cloudIntegrationRow) []duplicateGroup {
+	if len(rows) == 0 {
+		return nil
+	}
+
+	var groups []duplicateGroup
+	var current duplicateGroup
+	current.keeper = rows[0]
+
+	for i := 1; i < len(rows); i++ {
+		row := rows[i]
+		if row.AccountID == current.keeper.AccountID &&
+			row.Provider == current.keeper.Provider &&
+			row.OrgID == current.keeper.OrgID {
+			current.losers = append(current.losers, row)
+		} else {
+			groups = append(groups, current)
+			current = duplicateGroup{keeper: row}
+		}
+	}
+	groups = append(groups, current)
+
+	return groups
+}
+
+// mergeCloudIntegrationConfigs unions the EnabledRegions from all rows in the group into the keeper's config and updates
+func mergeCloudIntegrationConfigs(ctx context.Context, tx bun.Tx, group duplicateGroup) error {
+	regionSet := make(map[string]struct{})
+
+	// Parse keeper's config
+	parseRegions(group.keeper.Config, regionSet)
+
+	// Parse each loser's config
+	for _, loser := range group.losers {
+		parseRegions(loser.Config, regionSet)
+	}
+
+	// Build merged config
+	mergedRegions := make([]string, 0, len(regionSet))
+	for region := range regionSet {
+		mergedRegions = append(mergedRegions, region)
+	}
+
+	merged := cloudIntegrationAccountConfig{Regions: mergedRegions}
+	mergedJSON, err := json.Marshal(merged)
+	if err != nil {
+		return err
+	}
+
+	// Update keeper's config
+	_, err = tx.NewUpdate().
+		TableExpr("cloud_integration").
+		Set("config = ?", string(mergedJSON)).
+		Where("id = ?", group.keeper.ID).
+		Exec(ctx)
+	return err
+}
+
+// parseRegions unmarshals a config JSON string and adds its regions to the set.
+func parseRegions(configJSON string, regionSet map[string]struct{}) {
+	if configJSON == "" {
+		return
+	}
+
+	var config cloudIntegrationAccountConfig
+	if err := json.Unmarshal([]byte(configJSON), &config); err != nil {
+		return
+	}
+
+	for _, region := range config.Regions {
+		regionSet[region] = struct{}{}
+	}
+}