CI/CD Pipeline Throughput Simulator

Pipeline Load

Arrival rate 60/hr

Build workers 2

Test workers 2

Test failure rate 12%

Live Readouts

Throughput

0/hr

Avg pipeline time

0.0 min

Bottleneck stage

—

Total backlog

Legend

Waiting in queue

Building

Testing

Deploying

Test failed → retry

How it works

Each deployment job passes through three stages — Build, Test, Deploy — modeled as an M/M/c queue: jobs arrive as a Poisson process at rate λ, each stage has c parallel workers, and each worker's service time is drawn from an exponential distribution around a fixed mean (Build 2 min, Test 4 min, Deploy 1 min).

Arrivals:     inter-arrival ~ Exp(λ),  λ = rate / 60  (jobs/min)
Service:      duration ~ Exp(1/μ),   μ = 1 / mean-service-time
Utilization:  ρ = λ_stage / (c · μ)   (bottleneck = highest ρ)
Little's Law: L = λ · W   (jobs in system = arrival rate × time in system)

A stage with ρ close to or above 1 cannot keep up: its queue grows without bound, exactly what a real CI/CD pipeline looks like when reviewers merge faster than the test runners can execute. A failed test doesn't discard the job — it loops back to the Build queue as a retry, which raises the effective arrival rate at Build above the raw λ you set, often making Build the real bottleneck even though Test looks slower on paper.

Arrival rate — how often new commits request a deployment.
Build / Test workers — parallel capacity c at each stage; Deploy is fixed at 2 workers.
Test failure rate — probability a job fails testing and is sent back to Build.
Cube height stacked in a queue is the visual backlog — a tower that keeps growing means that stage is over capacity (ρ ≥ 1).

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.