feat(fleet): add service type enum and infra scope

Replace the free-form service_type string with a ServiceType enum (ccms,
ccmc, ccb, cces, ccsa, ccnc, ccem). The short code is the canonical value
stored in the DB, the config-tree directory name and a NATS discovery
subject token; Description() gives the human-readable name.

Add a second scope alongside per-node agents: ScopeInfra covers
cluster-independent monitoring infrastructure services, which are
identified by (hostname, service_type) and carry an empty cluster.
InfraRegistry is the ScopeInfra sibling of Registry and shares the same
table, state machine and config-revision handshake. It deliberately
exposes no StartSweep: MarkStale ages rows by last_heartbeat regardless
of scope, so a single sweep goroutine per process covers both scopes.

On the repository side ServiceDB gains a Scope column and the two
queries the discovery publisher and infra enumeration need, ListActive()
and ListByScope(). ResetConnection() now also resets the fleet
repository singleton so tests get a fresh handle.

Migration 13 is edited in place rather than superseded by a new
migration: the service table is unreleased, so no existing deployment
has it yet.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

internal/repository/dbConnection.go

@@ -136,6 +136,8 @@ func ResetConnection() error {
 	userRepoOnce = sync.Once{}
 	userCfgRepoInstance = nil
 	userCfgRepoOnce = sync.Once{}
+	fleetRepoInstance = nil
+	fleetRepoOnce = sync.Once{}
 
 	return nil
 }

internal/repository/fleet.go

@@ -52,11 +52,15 @@ type ServiceDB struct {
 	LastHeartbeat  sql.NullInt64  `db:"last_heartbeat"`
 	ConfigRevision int64          `db:"config_revision"`
 	MetaData       sql.NullString `db:"meta_data"`
+	// Scope is 'cluster' for per-node agents (the default) or 'infra' for
+	// cluster-independent monitoring infrastructure services. Infra rows carry
+	// an empty cluster.
+	Scope string `db:"scope"`
 }
 
 const namedServiceInsert string = `
-INSERT INTO service (cluster, hostname, service_type, instance_id, state, registered_at, config_revision, meta_data)
-	VALUES (:cluster, :hostname, :service_type, :instance_id, 'pending', :registered_at, 0, :meta_data);`
+INSERT INTO service (cluster, hostname, service_type, instance_id, scope, state, registered_at, config_revision, meta_data)
+	VALUES (:cluster, :hostname, :service_type, :instance_id, :scope, 'pending', :registered_at, 0, :meta_data);`
 
 // RegisterService upserts a service by (cluster, hostname, service_type): a
 // service registering for the first time is inserted with config_revision 0;
@@ -75,6 +79,7 @@ func (r *FleetRepository) RegisterService(svc *ServiceDB) (int64, error) {
 	case nil:
 		if _, uerr := sq.Update("service").
 			Set("instance_id", svc.InstanceID).
+			Set("scope", svc.Scope).
 			Set("state", "pending").
 			Set("registered_at", svc.RegisteredAt).
 			Set("meta_data", svc.MetaData).
@@ -156,13 +161,13 @@ func (r *FleetRepository) SetConfigRevision(instanceID string, revision int64) e
 
 var serviceColumns = []string{
 	"id", "cluster", "hostname", "service_type", "instance_id",
-	"state", "registered_at", "last_heartbeat", "config_revision", "meta_data",
+	"state", "registered_at", "last_heartbeat", "config_revision", "meta_data", "scope",
 }
 
 func scanService(row interface{ Scan(...any) error }) (*ServiceDB, error) {
 	svc := &ServiceDB{}
 	if err := row.Scan(&svc.ID, &svc.Cluster, &svc.Hostname, &svc.ServiceType, &svc.InstanceID,
-		&svc.State, &svc.RegisteredAt, &svc.LastHeartbeat, &svc.ConfigRevision, &svc.MetaData); err != nil {
+		&svc.State, &svc.RegisteredAt, &svc.LastHeartbeat, &svc.ConfigRevision, &svc.MetaData, &svc.Scope); err != nil {
 		return nil, err
 	}
 	return svc, nil
@@ -210,3 +215,56 @@ func (r *FleetRepository) ListByCluster(cluster string) ([]*ServiceDB, error) {
 	}
 	return services, rows.Err()
 }
+
+// ListActive returns all services currently in the 'active' state, ordered by
+// cluster then hostname. It is the roster source for the fleet discovery
+// publisher: only live services are advertised to peers.
+func (r *FleetRepository) ListActive() ([]*ServiceDB, error) {
+	rows, err := sq.Select(serviceColumns...).From("service").
+		Where("state = ?", "active").
+		OrderBy("cluster ASC", "hostname ASC").
+		RunWith(r.DB).Query()
+	if err != nil {
+		cclog.Errorf("Error while listing active services: %v", err)
+		return nil, err
+	}
+	defer rows.Close()
+
+	services := make([]*ServiceDB, 0)
+	for rows.Next() {
+		svc, err := scanService(rows)
+		if err != nil {
+			cclog.Warn("Error while scanning rows (ListActive)")
+			return nil, err
+		}
+		services = append(services, svc)
+	}
+	return services, rows.Err()
+}
+
+// ListByScope returns all services with the given scope ('cluster' or
+// 'infra'), ordered by hostname/service_type. Used to enumerate
+// cluster-independent monitoring infrastructure services, which carry an empty
+// cluster and therefore are not discoverable via ListByCluster.
+func (r *FleetRepository) ListByScope(scope string) ([]*ServiceDB, error) {
+	rows, err := sq.Select(serviceColumns...).From("service").
+		Where("scope = ?", scope).
+		OrderBy("hostname ASC", "service_type ASC").
+		RunWith(r.DB).Query()
+	if err != nil {
+		cclog.Errorf("Error while listing services for scope '%s': %v", scope, err)
+		return nil, err
+	}
+	defer rows.Close()
+
+	services := make([]*ServiceDB, 0)
+	for rows.Next() {
+		svc, err := scanService(rows)
+		if err != nil {
+			cclog.Warn("Error while scanning rows (ListByScope)")
+			return nil, err
+		}
+		services = append(services, svc)
+	}
+	return services, rows.Err()
+}

internal/repository/fleet_test.go

@@ -0,0 +1,111 @@
+// Copyright (C) NHR@FAU, University Erlangen-Nuremberg.
+// All rights reserved. This file is part of cc-backend.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+package repository
+
+import (
+	"testing"
+)
+
+func TestFleetRepository(t *testing.T) {
+	setup(t) // migrates a fresh temp DB to the current schema version
+	repo := GetFleetRepository()
+
+	// A per-node (cluster) agent and a cluster-independent infra service.
+	clusterSvc := &ServiceDB{
+		Cluster: "fritz", Hostname: "node01", ServiceType: "agent",
+		InstanceID: "iid-cluster", Scope: "cluster", RegisteredAt: 1000,
+	}
+	infraSvc := &ServiceDB{
+		Cluster: "", Hostname: "ms01", ServiceType: "metric-store",
+		InstanceID: "iid-infra", Scope: "infra", RegisteredAt: 2000,
+	}
+
+	if _, err := repo.RegisterService(clusterSvc); err != nil {
+		t.Fatal(err)
+	}
+	if _, err := repo.RegisterService(infraSvc); err != nil {
+		t.Fatal(err)
+	}
+
+	t.Run("ListByScope isolates infra from cluster", func(t *testing.T) {
+		infra, err := repo.ListByScope("infra")
+		noErr(t, err)
+		if len(infra) != 1 {
+			t.Fatalf("want 1 infra service, got %d", len(infra))
+		}
+		if infra[0].InstanceID != "iid-infra" || infra[0].Cluster != "" || infra[0].Scope != "infra" {
+			t.Fatalf("unexpected infra row: %+v", infra[0])
+		}
+
+		cluster, err := repo.ListByScope("cluster")
+		noErr(t, err)
+		if len(cluster) != 1 || cluster[0].InstanceID != "iid-cluster" {
+			t.Fatalf("unexpected cluster scope rows: %+v", cluster)
+		}
+	})
+
+	t.Run("ListByCluster sees only its cluster", func(t *testing.T) {
+		rows, err := repo.ListByCluster("fritz")
+		noErr(t, err)
+		if len(rows) != 1 || rows[0].InstanceID != "iid-cluster" {
+			t.Fatalf("unexpected ListByCluster(fritz): %+v", rows)
+		}
+	})
+
+	t.Run("scope round-trips through GetByInstanceID", func(t *testing.T) {
+		got, err := repo.GetByInstanceID("iid-infra")
+		noErr(t, err)
+		if got.Scope != "infra" || got.State != "pending" {
+			t.Fatalf("unexpected: scope=%q state=%q", got.Scope, got.State)
+		}
+	})
+
+	t.Run("heartbeat is instance-id keyed and scope-agnostic", func(t *testing.T) {
+		affected, err := repo.Heartbeat("iid-infra", 3000)
+		noErr(t, err)
+		if affected != 1 {
+			t.Fatalf("want 1 row affected, got %d", affected)
+		}
+		affected, err = repo.Heartbeat("does-not-exist", 3000)
+		noErr(t, err)
+		if affected != 0 {
+			t.Fatalf("want 0 rows affected for unknown instance, got %d", affected)
+		}
+	})
+
+	t.Run("ListActive returns only active services", func(t *testing.T) {
+		// iid-infra was activated by the heartbeat above; iid-cluster is still pending.
+		active, err := repo.ListActive()
+		noErr(t, err)
+		if len(active) != 1 || active[0].InstanceID != "iid-infra" || active[0].State != "active" {
+			t.Fatalf("unexpected ListActive result: %+v", active)
+		}
+	})
+
+	t.Run("config revision persists", func(t *testing.T) {
+		noErr(t, repo.SetConfigRevision("iid-infra", 42))
+		got, err := repo.GetByInstanceID("iid-infra")
+		noErr(t, err)
+		if got.ConfigRevision != 42 {
+			t.Fatalf("want config_revision 42, got %d", got.ConfigRevision)
+		}
+	})
+
+	t.Run("re-register keeps revision, resets scope and state", func(t *testing.T) {
+		// Same identity triple, but arriving as an infra registration again.
+		reReg := &ServiceDB{
+			Cluster: "", Hostname: "ms01", ServiceType: "metric-store",
+			InstanceID: "iid-infra-2", Scope: "infra", RegisteredAt: 4000,
+		}
+		if _, err := repo.RegisterService(reReg); err != nil {
+			t.Fatal(err)
+		}
+		got, err := repo.GetByInstanceID("iid-infra-2")
+		noErr(t, err)
+		if got.State != "pending" || got.Scope != "infra" || got.ConfigRevision != 42 {
+			t.Fatalf("unexpected after re-register: %+v", got)
+		}
+	})
+}

internal/repository/migration.go

@@ -21,7 +21,8 @@ import (
 // is added to internal/repository/migrations/sqlite3/.
 //
 // Version history:
-//   - Version 13: Service table (fleet service registration/heartbeat)
+//   - Version 13: Service table (fleet service registration/heartbeat), incl.
+//     scope column (cluster vs cluster-independent infra services)
 //   - Version 12: Sessions table (server-side sessions via alexedwards/scs)
 //   - Version 11: Optimize job table indexes (reduce from ~78 to 48, add covering/partial indexes)
 //   - Version 10: Node table

internal/repository/migrations/sqlite3/13_add-service-table.down.sql

@@ -1,3 +1,4 @@
 DROP INDEX IF EXISTS services_state_heartbeat;
+DROP INDEX IF EXISTS services_scope;
 DROP INDEX IF EXISTS services_cluster;
 DROP TABLE IF EXISTS "service";

internal/repository/migrations/sqlite3/13_add-service-table.up.sql

@@ -4,6 +4,8 @@ CREATE TABLE "service" (
     hostname VARCHAR(255) NOT NULL,
     service_type VARCHAR(255) NOT NULL,
     instance_id VARCHAR(64) NOT NULL,
+    scope VARCHAR(32) NOT NULL DEFAULT 'cluster'
+        CHECK (scope IN ('cluster', 'infra')),
     state VARCHAR(32) NOT NULL DEFAULT 'pending'
         CHECK (state IN ('pending', 'active', 'stale', 'deregistered')),
     registered_at INTEGER NOT NULL,
@@ -15,6 +17,7 @@ CREATE TABLE "service" (
 );
 
 CREATE INDEX IF NOT EXISTS services_cluster ON service (cluster);
+CREATE INDEX IF NOT EXISTS services_scope ON service (scope);
 CREATE INDEX IF NOT EXISTS services_state_heartbeat ON service (state, last_heartbeat);
 
 PRAGMA optimize;