Skip to content

Gke cluster autoscaler vs Gke compute classes

Gke cluster autoscaler is an open-source agent skill for cloud and infrastructure. Gke compute classes is an open-source agent skill for cloud and infrastructure.

  • Gke cluster autoscaler is best for why aren't my gke nodes scaling down after a traffic drop?.
  • Gke compute classes is best for set up a computeclass that runs batch pods on spot vms and falls back to on-demand nodes.
What it is
Gke cluster autoscaler is an open-source agent skill for cloud and infrastructure.
Gke compute classes is an open-source agent skill for cloud and infrastructure.
Best for
Why aren't my GKE nodes scaling down after a traffic drop?
Set up a ComputeClass that runs batch pods on Spot VMs and falls back to on-demand nodes
Kind
Skill
Skill
Pricing
Open source
Open source
Platforms
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
GitHub stars
24.4K
24.4K
Last commit
Nov 2026
Nov 2026
License
Apache-2.0
Apache-2.0
Open source
Good for
Why aren't my GKE nodes scaling down after a traffic drop?; Turn on cluster autoscaling for an existing GKE node pool; Diagnose pods stuck pending because of a zonal stockout
Set up a ComputeClass that runs batch pods on Spot VMs and falls back to on-demand nodes; Target L4 or H100 GPUs for an inference deployment through a ComputeClass; Debug a GKE pod that stays Pending with a noScaleUp event

What Gke cluster autoscaler does

  • Enable autoscaling settings

    Covers turning on Cluster Autoscaler and Node Auto Provisioning on older and newer GKE versions, plus manual node pool autoscaling limits.

  • Scale-down blocker scan

    Walks through the full list of pod, annotation, disruption budget and pool constraints that stop nodes from being removed, with a bash scanner script.

  • Autoscaler event tailing

    Includes a script that tails and parses autoscaler decisions from Cloud Logging visibility logs for a given cluster.

  • Stockout and quota handling

    Explains scale-up errors such as zonal stockouts, quota exceeded and IP exhaustion, and how location policies and fallback tiers respond.

  • Capacity buffers

    Describes the preview CapacityBuffer resource for keeping warm or suspended nodes ready to cut provisioning lag during traffic spikes.

  • Consolidation tuning

    Covers the optimize-utilization profile, consolidation delay settings and location policy choices for Spot and on-demand workloads.

What Gke compute classes does

  • Spot with on-demand fallback

    Shows how to order priority tiers so workloads prefer Spot capacity and fall back to on-demand nodes.

  • Accelerator and machine targeting

    Covers selecting GPUs, TPUs and specific machine families within a ComputeClass priority list.

  • Pending pod diagnosis

    Guides debugging of Pods stuck Pending or showing noScaleUp during node pool auto-creation.

  • Access restriction patterns

    Separates RBAC control of ComputeClass objects from admission policies and scale-up quotas that limit consumption.

  • Stockout and stateful guidance

    Explains zonal stockout cooldowns, fallback ladder design and dynamic-rwo storage for stateful workloads.

  • Autopilot mode on Standard

    Describes built-in autopilot ComputeClasses and when a custom class needs autopilot enabled.

Gke cluster autoscaler vs Gke compute classes: questions

Is Gke cluster autoscaler or Gke compute classes cheaper?

Both are open source. Compare current plans on each official site.

What is the difference between Gke cluster autoscaler and Gke compute classes?

Gke cluster autoscaler is an open-source agent skill for cloud and infrastructure. Gke compute classes is an open-source agent skill for cloud and infrastructure.

Keep comparing