Kubernetes Manifest Readiness Linter

Paste the manifests for one workload — Deployment, HorizontalPodAutoscaler, PodDisruptionBudget, whatever you have — and get the production-readiness review a person would give it. Nothing is uploaded: the parsing and the checks run in this page, in your browser.

Load an example

1 · Manifests

Paste everything for one workload, separated by ---. The more you include — HPA, PodDisruptionBudget, Service — the more the checks can see about how the pieces fit together.

Runtime

JVM workloads get extra checks — heap sizing and memory-based autoscaling behave differently.

Environment

Non-production downgrades the availability findings — single replicas and missing budgets matter less there.

Show findings at or above

2 · Report

What was parsed

Findings

Review comment

These checks encode conventions, not laws — a finding is a question worth answering, not an automatic defect. Some are deliberate (a hostPath mount for an APM socket, a CPU limit demanded by a noisy-neighbour policy), and the report says so where it can. It reads only what you paste, so it cannot see admission controllers, Kyverno or Gatekeeper policies, a service mesh, defaults injected by a mutating webhook, or values a Helm chart would supply. Nothing leaves the browser — the parser and every check run locally, so manifests with real hostnames and secret references are safe to paste.

What was parsed

Findings

Review commentCopy

Review comment