3D Staged Rollout Monitor | mysimulator.uk

Release

Regression severity 2.0x

Halt threshold 2.0x

Install base 250K

Stage dwell time 1.0 d

Clock

Stage

–

Status

–

Canary crash rate

–

vs baseline

–

On new version

–

Rollbacks

–

Sim time elapsed

–

How it works

Every point on the sphere is one instanced device from the install base. Points shift from stable (slate) to canary (blue) as the rollout stage advances; the core sphere's size and glow track the live canary share. Devices flash red when they throw a crash — a Poisson draw each simulated day, scaled by cohort population and the cohort's true crash rate.

canary_rate = baseline_rate × severity
halt when  canary_rate_observed > threshold × stable_rate_observed
           AND canary population ≥ min sample (500)
           AND condition sustained
on halt → rollback: canary% → 0, core flashes red
on healthy for dwell_time → advance to next stage ring

Regression severity — how much worse the new build's true crash rate is; 1x means the release is actually safe and the rollout should complete.
Halt threshold — how many times baseline the monitor tolerates before auto-halting; the four stage pillars light up as each checkpoint clears.
Install base — total devices; small bases leave too few canary devices at 1%/5% for the monitor to trust ("insufficient sample").
Stage dwell time — how long a stage must stay healthy before the rollout auto-advances.

Real-world relevance: the same staged-rollout / canary-analysis loop runs behind Google Play's staged rollout, the App Store's phased release, and deployment systems like Spinnaker or Argo Rollouts — catching a bad build on a few thousand devices instead of the whole install base.

Stable (old version) Canary (new version) Crash event

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