KubeStellar Console Knowledge Base
The Knowledge Base is the library behind KubeStellar Console’s AI Missions system. It contains 400+ ready-to-use mission prompts that install, configure, troubleshoot, and repair open source Kubernetes projects across your multi-cluster fleet — all powered by AI that saves you time and tokens.
Two Types of Mission Prompts
Every mission prompt in the Knowledge Base falls into of two classes:
Installation Missions
Installation missions walk you through deploying a CNCF or open source project on or more clusters. Each installation mission includes:
- Prerequisites — What needs to be in place before installing (Helm, specific CRDs, minimum Kubernetes version)
- Installation steps — Step-by-step commands with explanations
- Configuration options — Common configuration parameters and when to use them
- Verification — How to confirm the installation succeeded
- Uninstall steps — Clean removal instructions
- Upgrade path — How to upgrade to newer versions
Example: Installing Prometheus across a multi-cluster fleet
Mission: Install Prometheus
Type: install
Difficulty: beginner
Target: All clusters matching label "monitoring=enabled"
Steps:
1. Add the prometheus-community Helm repo
2. Create the monitoring namespace
3. Install kube-prometheus-stack with recommended values
4. Verify Prometheus pods are running
5. Confirm metrics scraping is active
The AI executes each step, adapting to your specific cluster configuration. If a step fails, the AI diagnoses the issue and suggests a fix before continuing.
Solution Missions
Solution missions are troubleshooting and repair prompts. They start from a symptom — “pods crashing”, “high latency”, “certificate expired” — and guide the AI through diagnosis and repair.
- Symptom description — What the problem looks like
- Diagnostic steps — Commands to gather information
- Root cause analysis — AI interprets the diagnostic output
- Repair actions — Specific fixes with approval gates
- Verification — Confirm the fix worked
- Resolution saving — Save the solution for future reuse
Example: Repairing certificate expiry across clusters
Mission: Fix Expired TLS Certificates
Type: repair
Difficulty: intermediate
Symptom: Services returning TLS handshake errors
Steps:
1. Scan all clusters for expired certificates
2. Identify which cert-manager issuers are affected
3. Trigger certificate renewal
4. Verify new certificates are propagated
5. Confirm services are healthy
How AI Missions Work
When you run a mission from the Knowledge Base, here’s what happens:
- You pick a mission — Browse the Knowledge Base or describe your problem in natural language
- AI reads the prompt — The mission prompt gives the AI context, steps, and expected outcomes
- AI adapts to your environment — The AI checks your actual cluster state (namespaces, versions, resources) and adjusts the steps
- Step-by-step execution — Each step runs with your approval. The AI shows what it will do before doing it
- Automatic error recovery — If a step fails, the AI diagnoses the failure and tries an alternative approach
- Resolution saved — After success, the resolution is saved to your personal or shared Knowledge Base for future reuse
Multi-Cluster Awareness
Unlike single-cluster tools, every Knowledge Base mission is multi-cluster aware:
- Cluster selection — Choose which clusters to target (by name, label, or “all”)
- Parallel execution — Run the same mission across multiple clusters simultaneously
- Drift detection — The AI identifies differences between clusters and adapts
- Fleet-wide verification — Confirm results across all targeted clusters, not just
CNCF Projects in the Knowledge Base
The Knowledge Base includes installation and solution missions for projects across all CNCF categories. Here’s what’s currently available:
Observability & Monitoring
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| Prometheus | Yes | 12+ troubleshooting prompts | Graduated |
| Grafana | Yes | 8+ troubleshooting prompts | — |
| Jaeger | Yes | 6+ troubleshooting prompts | Graduated |
| OpenTelemetry | Yes | 10+ troubleshooting prompts | Incubating |
| Fluentd | Yes | 5+ troubleshooting prompts | Graduated |
| Thanos | Yes | 7+ troubleshooting prompts | Incubating |
| Cortex | Yes | 4+ troubleshooting prompts | Incubating |
| Loki | Yes | 6+ troubleshooting prompts | — |
Networking & Service Mesh
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| Istio | Yes | 15+ troubleshooting prompts | Graduated |
| Envoy | Yes | 8+ troubleshooting prompts | Graduated |
| Cilium | Yes | 10+ troubleshooting prompts | Graduated |
| Calico | Yes | 7+ troubleshooting prompts | — |
| Linkerd | Yes | 6+ troubleshooting prompts | Graduated |
| CoreDNS | Yes | 5+ troubleshooting prompts | Graduated |
| NATS | Yes | 4+ troubleshooting prompts | Incubating |
Security & Policy
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| OPA / Gatekeeper | Yes | 8+ troubleshooting prompts | Graduated |
| Falco | Yes | 6+ troubleshooting prompts | Graduated |
| cert-manager | Yes | 10+ troubleshooting prompts | — |
| Kyverno | Yes | 7+ troubleshooting prompts | Incubating |
| Trivy | Yes | 5+ troubleshooting prompts | — |
| Vault | Yes | 8+ troubleshooting prompts | — |
Storage & Data
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| Longhorn | Yes | 6+ troubleshooting prompts | Incubating |
| OpenEBS | Yes | 5+ troubleshooting prompts | Sandbox |
| Rook / Ceph | Yes | 8+ troubleshooting prompts | Graduated |
| MinIO | Yes | 4+ troubleshooting prompts | — |
| Velero | Yes | 6+ troubleshooting prompts | — |
Application Delivery & GitOps
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| Argo CD | Yes | 10+ troubleshooting prompts | Graduated |
| Flux CD | Yes | 8+ troubleshooting prompts | Graduated |
| Helm | Yes | 12+ troubleshooting prompts | Graduated |
| Kustomize | Yes | 5+ troubleshooting prompts | — |
| Crossplane | Yes | 7+ troubleshooting prompts | Incubating |
Runtime & Orchestration
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| containerd | Yes | 6+ troubleshooting prompts | Graduated |
| Knative | Yes | 8+ troubleshooting prompts | Incubating |
| KEDA | Yes | 5+ troubleshooting prompts | Graduated |
| KubeVirt | Yes | 4+ troubleshooting prompts | Incubating |
| Volcano | Yes | 3+ troubleshooting prompts | Incubating |
Infrastructure & Provisioning
| Project | Install Mission | Solution Missions | Maturity |
|---|---|---|---|
| Cluster API | Yes | 6+ troubleshooting prompts | — |
| Metal3 | Yes | 4+ troubleshooting prompts | — |
| Tinkerbell | Yes | 3+ troubleshooting prompts | Sandbox |
| wasmCloud | Yes | 3+ troubleshooting prompts | Sandbox |
Resolution Knowledge Base
Beyond the pre-built missions, the Knowledge Base grows with every problem you solve. The Resolution Knowledge Base tracks successful troubleshooting outcomes so you never solve the same problem twice.
How It Works
- Complete a mission — After the AI successfully repairs an issue, you’re prompted to save the resolution
- Save with context — The resolution captures the symptom, diagnostic steps, root cause, and fix
- Personal or shared — Keep resolutions private or share with your organization
- Automatic matching — When a similar problem occurs, the Knowledge Base suggests matching past resolutions
- One-click reapply — Apply a previous resolution to a new occurrence with a single click
Personal vs Shared Resolutions
| Scope | Who Sees It | Best For |
|---|---|---|
| Personal | you | Custom fixes for your specific cluster configurations |
| Shared | Everyone in your organization | Common issues that affect the whole team |
Over time, the shared Knowledge Base becomes an institutional memory of how your organization operates Kubernetes — a living runbook that the AI can reference automatically.
Browsing the Knowledge Base
Open the Mission Browser from the console sidebar or the AI Missions panel.
Filtering and Search
- By mission type — Install, troubleshoot, repair, upgrade, deploy, analyze
- By CNCF category — Observability, networking, security, storage, runtime
- By maturity — Graduated, incubating, sandbox
- By difficulty — Beginner, intermediate, advanced
- By tags — Specific technologies, use cases, or symptoms
- Full-text search — Search across mission titles, descriptions, and steps
Mission Sources
Missions come from three sources:
| Source | Description |
|---|---|
| KubeStellar Community | Curated missions maintained by the KubeStellar team |
| GitHub Community | Missions from GitHub repos tagged with kubestellar-missions |
| Local | Your own imported or AI-generated missions |
Deep Linking
Every mission has a permanent URL you can share:
https://console.kubestellar.io/missions/install-prometheus
https://console.kubestellar.io/missions/fix-certificate-expiry
Share mission links in Slack, documentation, or runbooks to give your team instant access to proven solutions.
Creating Custom Missions
You can create your own missions and add them to the Knowledge Base:
- From natural language — Describe what you want to accomplish and the AI generates a mission
- From a template — Start from an existing mission and modify it
- From YAML/JSON — Import a mission definition file
- From a resolution — Convert a successful troubleshooting session into a reusable mission
Custom missions follow the same format as community missions and can be shared via the Marketplace or GitHub.
Why This Saves You Time and Tokens
Traditional Kubernetes troubleshooting means:
- Google the error message
- Read 5 Stack Overflow answers
- Try each suggestion manually
- Repeat across every affected cluster
With the KubeStellar Console Knowledge Base:
- AI matches the symptom to a known resolution
- Executes the fix across all affected clusters in parallel
- Verifies the fix worked everywhere
The result: What used to take hours of manual debugging now takes minutes of AI-guided resolution — across your entire multi-cluster fleet. And because mission prompts are optimized for the AI model, you use fewer tokens per resolution than free-form chat, saving both time and cost.