Struct CaPool 

pub struct CaPool { /* private fields */ }

Thread-safe pool of CA backend connections.

Routes enrollment requests to healthy CAs based on the configured FailoverStrategy. The pool is updated by the super::health::HealthChecker and read by request handlers concurrently.

Implementations

impl CaPool

pub fn new(connections: Vec<CaConnection>, config: PoolConfig) -> Self

Create a new pool with the given backends and configuration.

pub fn with_queue_config( connections: Vec<CaConnection>, config: PoolConfig, queue_config: QueueConfig, ) -> Self

Create a new pool with explicit queue-and-retry configuration.

pub fn select(&self) -> Option<CaConnection>

Select the best available CA for an enrollment request.

Returns None when no healthy CA is available and the fallback behavior is FallbackBehavior::Reject.

pub fn select_allowed(&self, allowed: &[String]) -> Option<CaConnection>

Select only within the caller’s authorized CA set.

pub fn record_success(&self, id: &CaId, latency: Duration)

Record a successful request to a CA, updating latency EMA.

pub fn record_failure(&self, id: &CaId)

Record a failed request, applying circuit-breaker logic (RHELBU-3536 R2).

pub fn should_reprobe(&self, id: &CaId) -> bool

Check whether a tripped circuit breaker should allow a re-probe.

pub fn set_health(&self, id: &CaId, state: HealthState)

Update the health state for a CA (called by the health checker).

pub fn status_snapshot(&self) -> HashMap<CaId, CaStatus>

Snapshot of current statuses for monitoring.

pub fn connections(&self) -> &[CaConnection]

All registered CA connections.

pub fn config(&self) -> &PoolConfig

Pool configuration.

pub fn is_queue_and_retry(&self) -> bool

Whether the fallback behavior is queue-and-retry.

pub fn enqueue_request(&self, csr_der: Vec<u8>) -> EnqueueResult

Enqueue a CSR for later retry when all CAs are down.

Drains expired entries before inserting. Returns EnqueueResult::Queued with a suggested Retry-After value on success, or EnqueueResult::QueueFull if the queue is at capacity.

pub fn drain_queue(&self) -> Vec<QueuedRequest>

Drain all non-expired queued requests for processing.

Called when a CA becomes available to process the backlog. Returns queued requests that have not exceeded the TTL.

pub fn queue_depth(&self) -> usize

Current number of queued requests (for monitoring).

pub fn queue_config(&self) -> &QueueConfig

Queue configuration.