Sink & Counter¶

Built-in result collectors for tracking events and latency.

Common reusable entities for simulations.

Provides ready-made Sink and Counter entities that eliminate the need to write boilerplate event-collecting classes in every example or test.

Sink ¶

Sink(name: str = 'Sink')

Bases: Entity

Event collector with latency tracking.

Computes per-event latency from context["created_at"]. When that key is missing the latency is recorded as 0.

Attributes:

Name	Type	Description
`events_received`	`int`	Total number of events handled.
`latencies_s`	`list[float]`	Per-event latency in seconds (same order as arrival).
`completion_times`	`list[Instant]`	Simulation `Instant` when each event arrived.

average_latency() -> float

Mean latency across all received events.

latency_time_series_seconds() -> tuple[
    list[float], list[float]
]

Return (completion_times_s, latencies_s) for plotting.

latency_stats() -> dict

Summary statistics over all recorded latencies.

Returns a dict with keys: count, avg, min, max, p50, p99.

Counter(name: str = 'Counter')

Bases: Entity

Simple event counter that tallies events by type.

Attributes:

Name	Type	Description
`total`	`int`	Total events received.
`by_type`	`dict[str, int]`	Mapping from `event_type` string to count.