Quick Start

From zero to running Fortem in your cluster. ~5 minutes.

Fortem v1 · K8s ≥ 1.26

⚠

Fortem v1 is in design partner access. The Helm chart and full documentation are shared with early partners after a discovery call. Book a 20-min call →

Prerequisites

A running Kubernetes cluster (EKS, GKE, AKS, or self-managed k8s ≥ 1.26). Helm 3.x installed. `kubectl` configured with cluster-admin permissions.

bash

# Verify cluster accesskubectl cluster-info# Verify Helmhelm version

Install via Helm

Add the Fortem Helm repository and install the operator into a dedicated namespace. The operator registers all CRDs and starts the AI engine.

bash

helm repo add fortem https://charts.fortem.devhelm repo updatehelm install fortem fortem/fortem \  --namespace fortem-system \  --create-namespace \  --set ai.provider=openai \  --set ai.apiKey=<YOUR_API_KEY># Installation takes ~3 minutes# Watch operator come up:kubectl get pods -n fortem-system -w

Access the Dashboard

Once the operator is running, port-forward or configure your Ingress. The dashboard auto-discovers all clusters, namespaces, and workloads — no manual configuration.

bash

# Quick access via port-forwardkubectl port-forward svc/fortem-ui 8080:80 -n fortem-system# Open: http://localhost:8080# Default credentials are printed in the install output.# Or configure Ingress:helm upgrade fortem fortem/fortem \  --namespace fortem-system \  --set ingress.enabled=true \  --set ingress.host=fortem.your-domain.com

Create your first Environment

In the AI Ops tab, describe what you need in plain English. Fortem generates a valid Environment CRD manifest, shows you a dry-run diff, and deploys after your approval.

bash

# Example — type this in the AI Ops input:"Create staging namespace for api-gateway with Postgres 15 and Redis"# Fortem generates:apiVersion: fortem.dev/v1alpha1kind: Environmentmetadata:  name: api-gateway-staging  namespace: api-gateway-staging  labels:    fortem.dev/managed: "true"    fortem.dev/ttl: 7dspec:  cluster: prod-eu-west  autoShutdown: true  workloads:    - kind: Deployment      name: api-gateway      image: api-gateway:latest      replicas: 1    - kind: StatefulSet      name: postgres      image: postgres:15      storage: {size: 10Gi, storageClass: gp3}    - kind: StatefulSet      name: redis      image: redis:7-alpine      storage: {size: 2Gi, storageClass: gp3}

AIOps — automatic optimisation

Once environments are running, Fortem's AIOps engine analyses CPU/memory usage, detects idle namespaces, identifies CrashLoopBackOff causes, and surfaces right-sizing recommendations. No configuration needed.

bash

# AIOps insights appear automatically in the dashboard.# Example alerts Fortem generates:[CRITICAL] OOMKilled 3× in 2h  → Deployment/worker · staging  → Memory limit 256Mi, peak 310Mi  → Recommended: increase limit to 512Mi[WARNING]  Idle namespace detected  → dev-pr-448: no traffic or deploys for 9 days  → Estimated savings: $180/mo if destroyed[INFO]     Right-size opportunity  → Deployment/auth-service · production  → CPU request 500m, p95 actual 48m  → Save $76/mo across 2 replicas

Key Concepts

Environment CRD

A Fortem custom resource that describes a full application stack (Deployments, StatefulSets, Services, Ingresses, PVCs). The operator reconciles it into native K8s objects.

fortem-system namespace

Where the Fortem operator, AI engine, and UI run. Isolated from your application workloads.

AI Engine

Powered by OpenAI, Anthropic, or Gemini (your choice). Used for NL→manifest generation, incident diagnosis, and AIOps recommendations. The API key stays in your cluster.

AIOps

Continuous background analysis: idle detection, right-sizing, OOM diagnosis, CrashLoop analysis, pending pod triage. Runs in your cluster — no data leaves.

autoShutdown

If set on an Environment, Fortem automatically scales down the namespace after the ttl period with no traffic or deploy activity.

ttl (Time-to-Live)

A label on Environment resources. After this period with no activity, Fortem suggests or automatically destroys the namespace, recovering cloud spend.

Ready to install Fortem in your cluster?

The Helm chart and full documentation are available to design partners. Book a 20-minute call — we'll walk through your setup and get you access.

Book a discovery call →

← Home