🔧 AI Nachrichten DistroWatch Weekly, Issue 1188(31.08.2026 um 03:21 Uhr)
🐧 Linux TippsDistribution Release: Grml 2026.09(04.09.2026 um 01:39 Uhr)
🔧 ProgrammierungDistribution Release: Talos Linux 1.14.0(04.09.2026 um 11:06 Uhr)
🐧 Linux TippsDistribution Release: Zenwalk GNU Linux Current-260905(05.09.2026 um 22:05 Uhr)
🔧 AI Nachrichten DistroWatch Weekly, Issue 1189(07.09.2026 um 02:18 Uhr)
🐧 Linux TippsSecurity: Denial of Service in grpcurl (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Denial of Service in syncthing (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Mehrere Probleme in darktable (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Mehrere Probleme in kamailio (Debian)(12.09.2026 um 00:28 Uhr)
🔧 AI Nachrichten DistroWatch Weekly, Issue 1188(31.08.2026 um 03:21 Uhr)
🐧 Linux TippsDistribution Release: Grml 2026.09(04.09.2026 um 01:39 Uhr)
🔧 ProgrammierungDistribution Release: Talos Linux 1.14.0(04.09.2026 um 11:06 Uhr)
🐧 Linux TippsDistribution Release: Zenwalk GNU Linux Current-260905(05.09.2026 um 22:05 Uhr)
🔧 AI Nachrichten DistroWatch Weekly, Issue 1189(07.09.2026 um 02:18 Uhr)
🐧 Linux TippsSecurity: Denial of Service in grpcurl (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Denial of Service in syncthing (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Mehrere Probleme in darktable (Fedora)(12.09.2026 um 00:28 Uhr)
🐧 Linux TippsSecurity: Mehrere Probleme in kamailio (Debian)(12.09.2026 um 00:28 Uhr)

🔧 Programmierung 🕛 vor 3 Monaten 5 Min Lesezeit
0

Stop Running LLM Workloads on Vanilla Kubernetes

↗ Quelle (dev.to)
🗣️ Stimme:
📑 Inhaltsübersicht

TL;DR: Kubernetes schedules LLM workloads well, but it does not give them the isolation boundary they need once they start calling tools, executing code, or handling tenant data.



Open Source Summit North America made one thing obvious: the cloud native crowd has moved from "can Kubernetes run LLM workloads?" to "what breaks when we trust Kubernetes too much?"



That is the right question.



The default Kubernetes security model assumes a pod is mostly an application packaging unit. It gives you namespaces, cgroups, seccomp, AppArmor, service accounts, admission control, and network policy. All of that matters. None of it changes the central fact that normal containers share the host kernel.



For a stateless API, that tradeoff is usually fine. For an LLM tool runner that can read files, call APIs, invoke Python, shell out to package managers, and chain actions across systems, that boundary starts looking thin.



The uncomfortable version is this: vanilla Kubernetes is orchestration, not containment.








CODE
apiVersion: node.k8s.io/v1
kind: RuntimeClass
metadata:
name: gvisor
handler: runsc
---
apiVersion: node.k8s.io/v1
kind: RuntimeClass
metadata:
name: kata
handler: kata






Then each workload declares the boundary it needs:




CODE
apiVersion: apps/v1
kind: Deployment
metadata:
name: tool-using-agent
spec:
replicas: 3
selector:
matchLabels:
app: tool-using-agent
template:
metadata:
labels:
app: tool-using-agent
spec:
runtimeClassName: kata
serviceAccountName: llm-agent
containers:
- name: agent
image: ghcr.io/example/tool-agent:2026.05






My rule of thumb:

































Workload Runtime Why
Plain inference API
runc or gvisor
Low tool risk, latency sensitive
Retrieval worker with narrow egress gvisor Better syscall boundary with less operational change
Agent that calls tools kata VM boundary per pod, Kubernetes friendly
Arbitrary code execution Firecracker style microVM Treat it like untrusted tenant compute


gVisor is the easiest first step because it integrates as an OCI runtime through runsc. Kata is the better fit when the isolation requirement is stronger and a VM per pod is acceptable. Firecracker is the most interesting boundary for code execution, but it is also the one I would least casually bolt onto an existing cluster without a real operations plan.






The Minimum Policy Set



The runtime is only one layer. I would not run LLM workloads without this set:




CODE
apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
name: llm-worker-egress
spec:
podSelector:
matchLabels:
app: tool-using-agent
policyTypes: ["Egress"]
egress:
- to:
- namespaceSelector:
matchLabels:
name: model-gateway
ports:
- protocol: TCP
port: 443
- to:
- namespaceSelector:
matchLabels:
name: telemetry
ports:
- protocol: TCP
port: 4317






Also make the service account boring:




CODE
apiVersion: v1
kind: ServiceAccount
metadata:
name: llm-agent
automountServiceAccountToken: false






If the workload does not need Kubernetes API access, do not mount a token. If it does, bind only the exact verbs it needs.






Benchmark Plan



I am not going to fake GPU numbers from a laptop. The package needs a real GPU node before publishing final performance claims.



This is the harness I would run:





  • Kubernetes Agent Sandbox:

  • gVisor:

  • Firecracker containerd: https://github.com/firecracker-microvm/firecracker-containerd

  • Vollständiger Original-Bericht
    Ausführliche Details, Code-Beispiele & Hersteller-Stellungnahme auf dev.to.
    ↗ Original-Artikel auf dev.to lesen
    Wie bewertest du diesen Beitrag?
    1 Klick Feedback
    Teilen mit Netzwerk & Team:

    Community-Analysen & Experten-Meinungen 0

    Verfasse deine eigene Analyse, teile Workarounds oder diskutiere diesen Vorfall im Blog.
    Noch keine Community-Analyse verfasst. Markiere einen Textabschnitt oder klicke oben auf Eigene Analyse verfassen“!
    Community Pulse: Relevanz-Einschätzung
    1 Klick Experten-Votum
    🔴 Akute Relevanz 0%
    🟡 In Evaluierung 0%
    🟢 Keine Auswirkung 0%
    Spannende Innovation 0%
    Verwandte Story-Cluster & Quellen (Vektor-KI)
    Port 8095 Engine
    1 Quelle
    9 Proofpoint alternatives. Pros & cons of the leading options
    1 Quelle
    What the DfE’s cyber security update means for multi-academy trusts
    1 Quelle
    Coffee with the Council Podcast: Celebrating 20 Years of Securing Payment Data
    Ähnliche Beiträge
    🔍 Verwandte News

    Auch interessante Nachrichten Stop Running LLM Workloads on Vanilla Kubernetes

    Thematisch verwandte Begriffe: Stop, Running, Workloads, Vanilla · 6 Treffer

    Laden...

    Videos werden geladen ...

    Laden...

    Beiträge werden geladen ...

    Laden...

    Videos werden geladen ...

    Laden...

    Beiträge werden geladen ...

    Laden...

    Videos werden geladen ...

    Laden...

    Beiträge werden geladen ...

    Laden...

    Videos werden geladen ...

    Laden...

    Beiträge werden geladen ...

    Laden...

    Videos werden geladen ...