· on-call: primary

— resiliency for real-time card payments

resiliency-engine · us-east-1simulated · click a cell to inject failure

auths/s · sim

—

dropped

last mttr · sim

—

healthy cells

3/3

inject failure →

Model of the production engine: detect p99 breach → reach quorum → shift Route 53 weights → drain queued auths to healthy cells. Timings compressed.

// 01 · service level

Production SLOs

window: trailing 12 months

// 02 · service registry

work, side projects, experiments

kubectl get svc

// 03 · runbook

RUNBOOK.md — last reviewed by operator

// 04 · on-call

escalation policy: email first