Engine overview

Create, operate, and inspect Firebolt Engines.

A FireboltEngine runs the compute nodes that serve queries. Every Engine references a ready FireboltInstance in the same namespace and may reference a FireboltEngineClass for shared defaults.

Use the Engine guides to operate the workload:

• Engine scaling covers replica changes, stopping and resuming, image updates, and rollout progress.
• Engine rollouts covers graceful and recreate strategies, readiness, drain checks, and capacity planning.
• Auto-stop and wake-up covers activity-based scale-down and Gateway-triggered wake-up.
• Object storage covers managed-table storage configuration.

For every field, phase, and condition, see the FireboltEngine CRD reference.

Inspect Engine status

kubectl get fireng -n firebolt

NAME        REPLICAS   PHASE    READY   GENERATION   AGE
my-engine   5          stable   True    2            24h

Use kubectl describe to inspect conditions and Kubernetes events:

kubectl describe fireng my-engine -n firebolt

An Engine is available when its phase is stable and its Ready condition is True. The transition phases creating, switching, draining, and cleaning indicate rollout progress. A phase of stopped means spec.replicas is 0 and the Engine is intentionally unavailable.

Worked manifests are available in the examples/ directory.