An agentic fleet intelligence platform for Rancher. AI diagnostics backed by 48,000+ real fixes, GPU management, Harvester HCI, and autonomous remediation with human-in-the-loop approval.
Install once. Get fleet management, operations tooling, Harvester HCI, GPU intelligence, and AI diagnostics — all inside the Rancher Dashboard.
Context-aware AI drawer that understands your page, cluster, and history. Diagnose issues and approve remediation actions directly from Rancher.
ShippingProactive issue detection across your fleet. Severity-ranked with recommended actions and one-click AI diagnosis.
ShippingUpgrade risk, CIS benchmarks, etcd backup, monitoring, support bundles, and audit log in a unified tabbed view.
ShippingFull VM lifecycle management. Nodes, networks, snapshots, images, backups, and upgrades with auto-balance recommendations.
ShippingFleet-wide inventory across NVIDIA H100, A100, AMD MI300X, Intel Flex. Utilization scoring and workload attribution.
ShippingFleet CRD support with error intelligence. Multi-cluster health overview and configuration drift detection.
ShippingNeuVector security posture and Longhorn storage health in a unified infrastructure view with correlated insights.
ShippingProject management with CPU, GPU, and memory quotas. Role-based access control and namespace governance.
ShippingEvery diagnostic links to the GitHub issue it came from. Confidence scores, match counts, and verifiable proof.
ProNot a chatbot. An agentic fleet orchestration layer that diagnoses issues, proposes remediations, and executes user-approved actions across your clusters via MCP — connecting infrastructure, ticketing, and observability in one intelligent loop.
Liz wraps a generic LLM. Pastures is an intelligence platform with a data moat and a full operations suite.
| Capability | SUSE Liz | Pastures | + Engine |
|---|---|---|---|
| Context-aware AIUnderstands current page and resource | ✓ | ✓ | ✓ |
| MCP agent routingSpecialized agents via MCP | ✓ | ✓ | ✓ |
| BYO model supportOpenAI, Anthropic, Ollama, Bedrock | ✓ | ✓ | ✓ |
| Curated fix corpus48K+ real resolutions from 18 repos | — | — | 48,000+ |
| Source citationsLinks to GitHub issues with confidence | — | — | ✓ |
| Fix verificationConfirms the fix actually worked | — | — | ✓ |
| Fleet correlationSame issue across multiple clusters | — | — | ✓ |
| Cluster memoryRemembers past issues per cluster | — | — | ✓ |
| GPU intelligenceH100, A100, MI300X, Intel Flex | — | ✓ | ✓ |
| Harvester HCIVMs, nodes, networks, snapshots, upgrades | — | ✓ | ✓ |
| Upgrade risk assessment | — | ✓ | ✓ |
| CIS Benchmark scanning | — | ✓ | ✓ |
| Requires Rancher Prime | Yes | No | No |
| Agentic fleet actionsAI proposes, human approves, system executes | ✓ | ✓ | ✓ |
| Production status | GA | Shipping | Shipping |
Pastures Engine indexes real issue resolutions — not documentation, not forums. Every fix was proven in production.
Bring your own model to get started. Connect to Pastures Engine when you need production-grade intelligence.
Pastures AI handles 90% of diagnostics. For the rest, you get direct access to engineers who know K3s, RKE2, Rancher, and Harvester inside out.
Priority response via dedicated Slack channel. Monthly cluster health review. Direct access to the engineering team.
Urgent production issue? Get a Kubernetes engineer on a call within minutes. No subscription required.
Named support engineer, custom SLAs, runbook development, on-call coverage, and quarterly architecture reviews.
Book a 30-minute demo. We'll show you what your clusters look like through Pastures.