Struct ClusterObserver 

pub struct ClusterObserver {
    pub node_id: u64,
    pub lifecycle: ClusterLifecycleTracker,
    pub topology: Arc<RwLock<ClusterTopology>>,
    pub routing: Arc<RwLock<RoutingTable>>,
    pub group_status: Weak<dyn GroupStatusProvider + Send + Sync>,
}

Aggregated observability handle for the cluster.

Stored in SharedState::cluster_observer as an Arc<ClusterObserver> so HTTP route handlers and the metrics endpoint can build snapshots without threading four separate handles through every call.

Construction is done exactly once — after start_cluster has returned and start_raft has built the RaftLoop. See nodedb::control::cluster::start_raft for the wiring.

Fields

node_id: u64

This node’s id.

lifecycle: ClusterLifecycleTracker

Lifecycle phase tracker (shared with start_cluster).

topology: Arc<RwLock<ClusterTopology>>

Shared cluster topology.

routing: Arc<RwLock<RoutingTable>>

Shared routing table.

group_status: Weak<dyn GroupStatusProvider + Send + Sync>

Type-erased per-group status provider backed by RaftLoop.

Held as a Weak on purpose: the owner of this observer (SharedState) is itself kept alive transitively by the RaftLoop, so a strong reference here would close a cycle that pins both forever and blocks clean shutdown. The loop is kept alive by its own spawned tasks; upgrade therefore succeeds throughout normal operation and only fails once the loop has been dropped on shutdown, where an empty status snapshot is correct.

Implementations

impl ClusterObserver

pub fn new( node_id: u64, lifecycle: ClusterLifecycleTracker, topology: Arc<RwLock<ClusterTopology>>, routing: Arc<RwLock<RoutingTable>>, group_status: Weak<dyn GroupStatusProvider + Send + Sync>, ) -> Self

pub fn snapshot(&self) -> ClusterInfoSnapshot

Build a complete ClusterInfoSnapshot for rendering via /cluster/status or /metrics.