skills.rest
sayalinvidia
sayalinvidiaCOMMUNITY
Explore all skills by this author →
GitHub

dynamo-troubleshoot

Diagnose and triage NVIDIA Dynamo Kubernetes deployment failures with read-only bundles.

Updated Apr 16, 2026
Search Tags:#debugging#troubleshooting#kubernetes#read-only#bundle#dynamo
One-click install
npx skills add https://github.com/sayalinvidia/sayali-skills-test --skill dynamo-troubleshoot
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: dynamo-troubleshoot
Source: https://github.com/sayalinvidia/sayali-skills-test/tree/main/skills/dynamo-troubleshoot
Command: npx skills add https://github.com/sayalinvidia/sayali-skills-test --skill dynamo-troubleshoot

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Diagnose failed or unhealthy Dynamo Kubernetes deployments by collecting read-only evidence, classifying failures, and guiding safe remediation.

Core Features & Use Cases

  • Read-only bundle collection for pods, events, and statuses to avoid exposing secrets.
  • Failure classification into categories like cluster/namespace, secret/model access, PVC/storage, image/runtime, GPU scheduling, DynamoGraphDeployment/operator, frontend/router, and endpoint/API.
  • Actionable troubleshooting guidance including strongest signal, likely cause, and smallest reversible fix.

Quick Start

Run the dynamo-troubleshoot skill against a target namespace to collect a read-only bundle and start diagnosing the issue.

Frequently Asked Questions about dynamo-troubleshoot

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I troubleshoot a failing Dynamo Kubernetes deployment?▼

To troubleshoot a failing Dynamo Kubernetes deployment, run a diagnostic process that collects a read-only evidence bundle of pod statuses and events, classifies the failure, and identifies the smallest reversible fix.

What is the safest way to diagnose frontend crashloops in a Dynamo Kubernetes environment?▼

Diagnosing frontend crashloops safely involves collecting a read-only evidence bundle that avoids exposing secrets, allowing you to classify the failure and identify the strongest signal without modifying cluster state.

How do I fix a model-download PVC failing in Dynamo without causing disruptions?▼

Fixing a failing model-download PVC requires extracting the strongest signal from a read-only evidence bundle and applying the smallest reversible fix to resolve the storage access issue safely.

What categories of failures can be classified when diagnosing DynamoGraphDeployment reconciliation issues?▼

When diagnosing DynamoGraphDeployment reconciliation issues, failures are classified into categories including cluster/namespace, secret/model access, PVC/storage, image/runtime, GPU scheduling, and operator errors.

Does Dynamo troubleshooting require cluster write access to resolve DGD reconciliation problems?▼

Dynamo troubleshooting does not require cluster write access; it collects a read-only bundle to avoid exposing secrets and proposes the smallest reversible fix to safely address DGD reconciliation problems.

Related Skills

Explore More Software Engineering Skills→
ObolNetworkObolNetwork
10

obol-stack

Monitor Obol Stack Kubernetes pods, logs, services, and deployments via the Kubernetes API.

Official
Intermediate
indrasvatindrasvat
3

rollout-check

Analyze Kubernetes deployment rollout status, pod readiness, events, and HPA state.

Community
Advanced
alastor0325alastor0325

check-log

Analyze Firefox log files to diagnose media, EME, CDM, and GPU crash issues.

Community
Intermediate
scitixscitix
225

pvc-debug

Diagnose PersistentVolumeClaim failures in Kubernetes using kubectl.

Official
Intermediate