~/DOCS/

How It Works

Components

Architect deploys four components into your cluster:

  • Admission controller: configures managed containers on pod creation
  • Control plane: coordinates checkpoint transfers during migrations
  • Daemon: per-node agent that orchestrates hibernation and wake
  • Shim: performs checkpoint/restore via CRIU (runc-architect) or gVisor (runsc-architect)
Each Node Cluster Your Pods Architect Daemon Kubernetes API Admission Controller Control Plane

Pod Startup

When a pod starts, Architect checks for an available checkpoint:

Yes Fails No runc-architect Yes No Fails runsc-architect Pod starts start-from-persistent-checkpointannotation set? Restore from PersistentCheckpoint Start fresh Runtime class? Checkpointavailable? Restore from checkpoint

The start-from-persistent-checkpoint annotation works for both runtime classes and, when set, is the only source of checkpoint data: on any failure the pod starts fresh. Without the annotation, runc-architect pods find checkpoints from other pods with the same template hash (this is how migration works); runsc-architect pods always start fresh.

Scale-Down

When a container has been idle for the configured duration:

Checkpoint and stop container Buffer incoming packets Reduce pod resources to zero Status: SCALED_DOWN Your Pod Architect

Scale-Up

By default only kubectl exec wakes a container. To wake on network traffic, enable network-monitor.

When a wake trigger arrives (kubectl exec or network packet):

Request arrives Restore from checkpoint Release buffered packets Restore original resources Ready Status: RUNNING User / Traffic Architect Your Pod

Migration

When a pod is deleted (e.g., node drain), Architect transfers state to the replacement pod:

Store checkpoint Find checkpoint Transfer checkpoint Restore container Replacement pod scheduled Node 1 Kubernetes API Node 2