Struct ServiceInstance 

pub struct ServiceInstance {
    pub service_name: String,
    pub spec: ServiceSpec,
    /* private fields */
}

Service instance manages a single service’s containers

Fields

service_name: String

spec: ServiceSpec

Implementations

impl ServiceInstance

pub fn new( service_name: String, spec: ServiceSpec, runtime: Arc<dyn Runtime + Send + Sync>, overlay_manager: Option<Arc<RwLock<OverlayManager>>>, ) -> Self

Create a new service instance

pub fn with_proxy( service_name: String, spec: ServiceSpec, runtime: Arc<dyn Runtime + Send + Sync>, overlay_manager: Option<Arc<RwLock<OverlayManager>>>, proxy_manager: Arc<ProxyManager>, ) -> Self

Create a new service instance with proxy manager for health-aware load balancing

pub fn set_node_id(&mut self, node_id: u64)

Set the local cluster node id. Used by ServiceManager to thread Cluster::node_id() down to container construction so new ContainerIds carry the owning node identity. Defaults to 0 (the single-node sentinel) when unset.

pub fn role_for_replica(&self, replica_idx: u32) -> String

Derive the replica group role for a 1-based replica_idx.

When spec.replica_groups is unset, returns "default" (the implicit single-group case). Otherwise walks groups in declaration order, accumulating each group’s count until replica_idx falls within the current group’s range, and returns that group’s role.

Replicas beyond the declared total fall back to "default".

pub fn with_dns(self, dns_server: Arc<DnsServer>) -> Self

Builder method to add DNS server for service discovery

pub fn set_dns_server(&mut self, dns_server: Arc<DnsServer>)

Set the DNS server for service discovery

pub fn set_proxy_manager(&mut self, proxy_manager: Arc<ProxyManager>)

Set the proxy manager for health-aware load balancing

pub fn set_health_states( &mut self, states: Arc<RwLock<HashMap<String, HealthState>>>, )

Set the shared health states map so health callbacks can bridge state back to ServiceManager

pub async fn last_pulled_digest(&self) -> Option<String>

Get the last observed image digest (after the most recent successful pull). Returns None when no pull has happened yet, when the runtime does not expose digests, or when no matching ImageInfo was found.

pub async fn scale_to(&self, replicas: u32) -> Result<()>

Scale to the desired number of replicas

This method uses short-lived locks to avoid blocking concurrent operations. I/O operations (pull, create, start, stop, remove) are performed without holding the containers lock to allow other operations to proceed.

Errors

Returns an error if image pull, container creation, or container lifecycle operations fail.

pub async fn replica_count(&self) -> usize

Get current number of replicas

pub async fn container_ids(&self) -> Vec<ContainerId>

Get all container IDs

pub async fn container_infos(&self) -> Vec<ContainerInfo>

Get per-container info (id, image, state, pid, overlay IP) for every live container in this instance.

Surfaces the REAL image reference each container was created from and its REAL lifecycle state (lowercased via ContainerState::as_str) so the API/ps no longer reports a hardcoded "running" with no image.

pub fn containers(&self) -> &RwLock<HashMap<ContainerId, Container>>

Get read access to the containers map

This allows callers to access container overlay IPs and other metadata without copying the entire map.

pub fn has_overlay_manager(&self) -> bool

Check if this service instance has an overlay manager configured

pub fn has_proxy_manager(&self) -> bool

Check if this service instance has a proxy manager configured

pub fn proxy_manager(&self) -> Option<&Arc<ProxyManager>>

Get the proxy manager for this instance, if configured.

pub fn has_dns_server(&self) -> bool

Check if this service instance has a DNS server configured