Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Diagnosing a TPU training throughput drop on GKE v1.0.0

Use this resource to diagnose a throughput or duty cycle drop in TPU training on GKE

New

  • 320 downloads
  • 205 likes
  • 178 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install tpu-training-slowdown-diagnosis

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent queries ML runs and monitored events, then correlates the drop with one-minute system metrics and node topology. It routes the diagnosis to workload bottlenecks, such as HBM capacity or host CPU, or to infrastructure throttling. You get the likely cause and the next checks.

Mostly picked by

Updated November 2, 2026

Card for the “Diagnosing a TPU training throughput drop on GKE” skill

Details

Generates with
Your agent's model
Licence
Apache-2.0

Similar resources

View all

What you get

  • SKILL.md: queries on ML runs and monitored events, with routing to the right branch
  • references/path-a-workload-bottlenecks.md: workload bottlenecks
  • references/path-b-infrastructure-throttling.md: infrastructure throttling

Good fit for

  • TPU training throughput drops without pods crashing
  • PERFORMANCE_DEGRADATION events fire during a training run

Not for

    Tags

    • tpu
    • performance
    • training
    • gke

    About

    You say

    “My TPU training duty cycle dropped 20% this week: find out whether the workload or the infrastructure is the cause.”

    Card for the “Diagnosing a TPU training throughput drop on GKE” skill

    Skill published by google under the Apache-2.0 license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to diagnose a throughput or duty cycle drop in TPU training on GKE
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE11.3 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/attribution.md2 KB
    • references/path-a-workload-bottlenecks.md2.3 KB
    • references/path-b-infrastructure-throttling.md2.3 KB
    • references/path-c-no-detected-analyzer.md1.5 KB
    • references/path-d-healthy-telemetry.md0.3 KB
    • SKILL.md14.4 KB

    Origin and licence

    Listed unmodified

    Author
    google
    Source
    github.com/google/skills/tree/81b0e8f8f40513c31b3388e600506e8b48a85bc2/skills/cloud/gke-ai-troubleshooting-tpu-performance-degradation
    Licence
    Apache-2.0

    Notice

    « gke-ai-troubleshooting-tpu-performance-degradation » (google, Apache-2.0) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.