controller

type AgentHealthController struct {
	// contains filtered or unexported fields
}

type AgentHealthControllerConfig struct {
	// Interval is how often to run the health check.
	// Default: 30 seconds.
	Interval time.Duration

	// StaleTimeout is how long since last heartbeat before marking an agent as offline.
	// Default: 90 seconds (1.5x the typical heartbeat interval of 60s).
	StaleTimeout time.Duration

	// Logger for logging.
	Logger *logger.Logger
}

type AuditRetentionController struct {
	// contains filtered or unexported fields
}

type AuditRetentionControllerConfig struct {
	// Interval is how often to run the retention check.
	// Default: 24 hours (once a day).
	Interval time.Duration

	// RetentionDays is how long to keep audit logs.
	// Logs older than this will be deleted.
	// Default: 365 days (1 year).
	RetentionDays int

	// BatchSize is the maximum number of logs to delete in one batch.
	// This prevents long-running transactions.
	// Default: 10000.
	BatchSize int

	// DryRun if true, only counts logs that would be deleted without actually deleting.
	// Useful for testing retention policies.
	// Default: false.
	DryRun bool

	// Logger for logging.
	Logger *logger.Logger
}

type Controller interface {
	// Name returns the unique name of this controller.
	Name() string

	// Interval returns how often this controller should run.
	Interval() time.Duration

	// Reconcile performs the reconciliation logic.
	// It should be idempotent - running multiple times should have the same effect.
	// Returns the number of items processed and any error encountered.
	Reconcile(ctx context.Context) (int, error)
}

type JobRecoveryController struct {
	// contains filtered or unexported fields
}

type JobRecoveryControllerConfig struct {
	// Interval is how often to run the job recovery check.
	// Default: 60 seconds.
	Interval time.Duration

	// StuckThresholdMinutes is how long a job can be in acknowledged/running state
	// without progress before being considered stuck.
	// Default: 30 minutes.
	StuckThresholdMinutes int

	// TenantStuckThresholdMinutes is how long a tenant command can be assigned
	// to an agent without being picked up before being reassigned.
	// Default: 10 minutes (shorter than platform jobs as tenant agents poll more frequently).
	TenantStuckThresholdMinutes int

	// MaxRetries is the maximum number of retry attempts for a job.
	// After this many retries, the job will be marked as failed.
	// Default: 3.
	MaxRetries int

	// MaxQueueMinutes is how long a job can wait in the queue before expiring.
	// Default: 60 minutes.
	MaxQueueMinutes int

	// Logger for logging.
	Logger *logger.Logger
}

type Manager struct {
	// contains filtered or unexported fields
}

type ManagerConfig struct {
	// Metrics collector (optional)
	Metrics Metrics

	// Logger (required)
	Logger *logger.Logger
}

type Metrics interface {
	// RecordReconcile records a reconciliation run.
	RecordReconcile(controller string, itemsProcessed int, duration time.Duration, err error)

	// SetControllerRunning sets whether a controller is running.
	SetControllerRunning(controller string, running bool)

	// IncrementReconcileErrors increments the error counter.
	IncrementReconcileErrors(controller string)

	// SetLastReconcileTime sets the last reconcile timestamp.
	SetLastReconcileTime(controller string, t time.Time)
}

type NoopMetrics struct{}

type PrometheusMetrics struct {
	// contains filtered or unexported fields
}

type QueuePriorityController struct {
	// contains filtered or unexported fields
}

type QueuePriorityControllerConfig struct {
	// Interval is how often to recalculate queue priorities.
	// Default: 60 seconds.
	Interval time.Duration

	// Logger for logging.
	Logger *logger.Logger
}