Struct NodeMonitor 

pub struct NodeMonitor { /* private fields */ }

Tracks node health based on heartbeat records and configurable timeouts.

Implementations

impl NodeMonitor

pub fn new(timeout_ms: u64) -> Self

Create a new monitor.

• timeout_ms — a node is considered dead when no heartbeat has been received for this many milliseconds.

pub fn with_max_history(self, max_history: usize) -> Self

Set the maximum heartbeat history to retain per node.

pub fn register(&mut self, node: NodeInfo)

Register a new node. Overwrites any previous registration for the same ID.

pub fn record_heartbeat( &mut self, node_id: &str, received_at: u64, latency_ms: u64, ) -> bool

Record a heartbeat received from a node.

Sets the node’s state to Alive and appends the record to its history. Returns false if the node is not registered.

pub fn check_timeouts(&mut self, now: u64) -> Vec<String>

Scan all registered nodes and mark those whose last heartbeat is older than timeout_ms as Dead.

Returns the IDs of nodes that transitioned to Dead in this call.

pub fn state(&self, node_id: &str) -> Option<&NodeState>

Return the current state of a node.

pub fn alive_nodes(&self) -> Vec<&NodeInfo>

Return references to all nodes whose state is Alive.

pub fn dead_nodes(&self) -> Vec<&NodeInfo>

Return references to all nodes whose state is Dead.

pub fn avg_latency(&self, node_id: &str) -> Option<f64>

Compute the average heartbeat latency for a node.

Returns None if the node has no heartbeat history.

pub fn node_count(&self) -> usize

Total number of registered nodes.

pub fn remove(&mut self, node_id: &str) -> bool

Deregister a node. Returns true if it existed.

pub fn last_heartbeat(&self, node_id: &str) -> Option<&HeartbeatRecord>

Return the most recent heartbeat record for a node, if any.

pub fn heartbeat_count(&self, node_id: &str) -> usize

Number of heartbeat records stored for a node.