This page lists all requirements for deploying Jitera Self-Hosted. Items are categorized as Mandatory (deployment will not function without them) or Optional (enhance functionality but can be added later).Documentation Index
Fetch the complete documentation index at: https://docs.jitera.ai/llms.txt
Use this file to discover all available pages before exploring further.
Mandatory Requirements
Kubernetes Cluster
Jitera Self-Hosted is verified with Kubernetes 1.35.x. The following environments are verified by Jitera:| Platform | Type |
|---|---|
| Amazon EKS | Managed |
| Azure AKS | Managed |
| On-premises (kubeadm, k3s, RKE2) | Self-managed |
Earlier versions (1.28-1.34) may work but are either in Extended Support on AWS EKS (with premium pricing) or require the LTS support plan on Azure AKS (Premium tier). If omitting the version flag on cluster creation commands, the cloud provider selects the latest stable version automatically.
Cluster Specifications
| Component | Production Minimum | Evaluation Minimum |
|---|---|---|
| CPU architecture | x86-64 (AMD64) — ARM is not supported | Same |
| CPU (total) | 16 cores | 12 cores |
| Memory (total) | 64 GB | 48 GB |
| Storage (per node) | 200 GB SSD | 200 GB SSD |
The evaluation minimum (e.g., 3x
m6i.xlarge on AWS or 3x Standard_D4s_v3 on Azure) is sufficient for development, proof-of-concept, and small-team deployments (up to 10 users). For production workloads, use the production minimum or refer to the Sizing Guide for detailed tier recommendations.Required Software
| Tool | Minimum Version |
|---|---|
| Helm | 3.20.x (verified) |
| kubectl | 1.35.x (verified) |
Helm 4.x is not supported. Use Helm 3.20.x.
The Helm chart includes Kong as an ingress controller. You do not need to install a separate ingress controller (e.g., NGINX) unless your environment requires one for other workloads.
Jitera Registry Credentials
Access to Jitera container images requires authentication credentials. These will be provided by Jitera after the contract is signed. See Registry Credentials for configuration details.Jitera Helm Chart
The Jitera Helm chart is provided as a zip file. It will be delivered by Jitera along with the registry credentials after the contract is signed.LLM Provider
Jitera requires an LLM provider for AI-powered code generation, chat, and document analysis. One of the following is required as the primary provider:| Provider | Configuration |
|---|---|
| OpenAI | API key with access to GPT-4 class or later models |
| Azure OpenAI | Azure OpenAI resource with model deployments |
gpt-4.1) and an embeddings model (e.g. text-embedding-ada-002) must be available from your chosen provider.
See AI Configuration for provider setup, model requirements, and additional provider options.
DNS Hostnames
Two resolvable hostnames belonging to the same domain are required:| Hostname | Purpose | Example |
|---|---|---|
| Main domain | Application UI, API, and services | app.example.com |
| Chat domain | Chat/AI interface (SWEF service) | chat.example.com |
example.com) because authentication cookies are set on the root domain.
Additional hostnames are required if you enable optional services: MinIO API (
minio.example.com), MinIO Console (minio-console.example.com), and Grafana (grafana.example.com). These should also resolve to the load balancer IP.Server Certificates
TLS certificates are required for HTTPS access. The certificate must cover both hostnames (main domain and chat domain). The following methods are verified:| Method | Notes |
|---|---|
| cert-manager | Automatic certificate management with Let’s Encrypt or other ACME CAs |
| AWS ACM | AWS Certificate Manager, used with AWS Load Balancer |
Object Storage
Jitera requires object storage for file uploads, generated code artifacts, and project exports. The following providers are verified:| Provider | Type |
|---|---|
| AWS S3 | Managed |
| Azure Blob Storage | Managed |
| MinIO (in-cluster) | Self-hosted, included in Helm chart |
SMTP Service
An SMTP service is required for email notifications, user invitations, and password resets. The following are verified:| Provider | Type |
|---|---|
| AWS SES | Managed |
| Azure Communication Services | Managed |
| SendGrid | Managed |
| Postfix (in-cluster) | Self-hosted, included in Helm chart |
When using SendGrid, set
mailer.smtp_settings.user_name to apikey (literal string) and use your SendGrid API key as the password.Optional Requirements
GitHub Integration
A GitHub App is required to connect GitHub repositories as code sources within Jitera projects. See GitHub Integration for setup instructions.GitLab Integration
A self-managed GitLab instance (version 16.8+) is required to connect GitLab repositories as code sources. A Personal Access Token with admin-level access is configured through the Jitera UI. See GitLab Integration for setup instructions.External Services
All data stores, caches, and message brokers are bundled in the Helm chart and deployed in-cluster by default. No external setup is required for initial deployment.| Service | Type |
|---|---|
| PostgreSQL | Database |
| PGVector | Database |
| MongoDB | Database |
| Redis | Cache |
| RabbitMQ | Message broker |
Additional AI Providers
Beyond the mandatory primary LLM provider, you can configure additional providers for expanded model access:| Provider | Purpose |
|---|---|
| AWS Bedrock | Access to Claude, Llama, and other models via Bedrock |
| Anthropic | Direct API access to Claude models |
| Google AI | Access to Gemini models |
Web Search Agent
To enable web search, URL reading, and deep research capabilities in Boost, configure one of the following search backends:| Backend | Configuration |
|---|---|
| Tavily (recommended) | API key — no self-hosted infrastructure needed |
| SearXNG | Self-hosted metasearch engine URL |
r.jina.ai) must be reachable for URL-to-text conversion. See AI Configuration — Web Search Agent for details.
Monitoring Stack
A full monitoring stack (Prometheus, Grafana, Loki, OpenTelemetry) is included in the Helm chart and enabled by default. It can be disabled if you use an existing monitoring solution.Network Requirements
For detailed firewall rules, port requirements, and network access patterns, see Network and Firewall.Checklist
Mandatory
- Kubernetes cluster running (1.35.x) with Helm 3.20.x and kubectl installed
- Jitera registry credentials received
- Jitera Helm chart zip file received
- LLM provider configured (OpenAI API key or Azure OpenAI resource)
- Two DNS hostnames configured on the same domain
- Object storage ready (S3, Azure Blob, or MinIO)
- SMTP service configured
- TLS certificates ready (cert-manager or AWS ACM)
Optional
- GitHub OAuth application created (for repository integration)
- GitLab Personal Access Token ready (for repository integration)
- Additional AI provider keys (Bedrock, Anthropic, Google AI)
- Web search backend configured (Tavily API key or SearXNG URL)
- External services configured (for production HA)
What’s Next?
AWS EKS
Deploy on Amazon EKS
Azure AKS
Deploy on Azure AKS
On-premises
Deploy on self-managed Kubernetes