Queue operations / deterministic seed 361

Backpressure Arcade

Keep a finite queue useful while Poisson arrivals, exponential service, client deadlines, retries, and delayed workers all compete for the same capacity.

System state STABLE Capacity is ahead of admitted load.

Live system boundary

Arrival gate → finite FIFO → worker pool

T+00:00.0 6.0 offered/s · 8.4 capacity/s
The live queue visualization requires canvas. All measured values are also available in the metric cards below. running / seeded
Steady load constant profile
arrival waiting timed out work completion worker booting

Throughput

0.00/s server completions, trailing 10s

Utilization

0% busy worker-time / ready worker-time

Queue depth

0 / 32 FIFO waiting only; service is separate

Drops

0 0.0% of offered attempts

Retry amplification

1.00× attempts created / original requests

p50 wait

0 ms queue admission to service start, trailing 60s

p95 wait

0 ms tail queue wait, trailing 60s

SLA

100% logical requests within original deadline

Workers

4 / 4 ready / desired; booting and draining are real

Arcade score

1000 SLA + useful work − drops, retries, and tail wait

Pressure trace

Actual samples from the last 60 simulated seconds.

queue throughput capacity
Queue depth, throughput, and service capacity history A sixty-second trace generated from the running simulation. high zero −60s now

Measured Little’s Law

Same boundary, since reset.

0.00 ≈ 0.00 × 0.00 = 0.00
L̄: mean in system0.00 jobs
λ: departure rate0.00 /s
W̄: mean sojourn0.00 s
Completed cohort0 jobs

Boundary gap: 0.0%

The gap is expected during transients: L̄ includes work still inside at the right boundary, while W̄ is only known for completed attempts. It should tighten in a stable, long-running system.

Operations feed

  1. Normal preset armed with deterministic seed 36101.