Lifecycle
IMP provides snapshot-based lifecycle primitives for save/restore and fast VM provisioning.
Sections
Concepts
- ImpVMSnapshot: captures a running VM's memory and disk state. Supports one-shot and scheduled recurring captures.
- ImpWarmPool: keeps a pool of pre-booted VMs ready for instant assignment. Pool members boot from an elected snapshot base.
Warm pools are the recommended path for latency-sensitive workloads — VMs are booted and waiting before they are needed.
Runtime and Pressure Handling
imp-runtimeusesupdateStrategy: OnDelete— it never rolls automatically. See Day 0 — Runtime OnDelete for cordon/drain and delete-pod rollout.pressureLifecycle.enabledisfalseby default (opt-in). When enabled, the operator suspends largest VMs first onMemoryPressureand never auto-resumes. See Day 0 — pressureLifecycle and Architecture — Operational Notes.
