Skip to main content
Nebuly supports self-hosted deployments, giving you full control over your infrastructure and data. You deploy and manage the entire platform within your own cloud environment, ensuring that all your data — including traces, logs, and analytics — remains within your infrastructure.

Use cases

Self-hosting is designed for organizations with specific requirements:
  • Data residency and compliance: Meet regulatory or contractual obligations by keeping all data within your own cloud account and region.
  • Security posture and isolation: Deploy the platform behind your firewall or VPN, using your own IAM policies, encryption keys, and audit trails. This ensures sensitive data never traverses external networks.
  • Access to private resources: Connect to internal LLM models, proprietary tools, or private APIs that are not accessible from the public internet. The platform runs within your network and can access resources in your VPC or private network.

How it works

Nebuly deploys as a single unified platform within your infrastructure. You own and operate the entire stack. The deployment process has two steps:
  1. Provision infrastructure using the Nebuly Terraform module for your cloud provider. This creates all required cloud resources (for details, look at the Terraform module for your cloud provider) and outputs a set of Helm values.
  2. Install the platform using the Nebuly Helm chart, passing the values produced by Terraform.
All data — including traces, logs, experiment records, and analytics — is stored exclusively within your infrastructure. By default, only licensing data and infrastructure health checks are exported outside your network; no sensitive data ever leaves your environment. Nebuly does not have access to your deployment unless you explicitly grant it for support purposes.

Deployment options

Nebuly provides official Terraform modules for self-hosting on major cloud providers:
  • AWS: Terraform with EKS and Helm
  • GCP: Terraform with GKE and Helm
  • Azure: Terraform with AKS and Helm
Nebuly requires using these Terraform modules because they are kept up-to-date with best practices, minimize configuration issues, and ensure Nebuly can efficiently troubleshoot performance and operational issues if needed.

Shared responsibility

When you self-host, uptime is a shared responsibility between your team and Nebuly:
  • Nebuly is responsible for responding quickly when you have issues, collaboratively resolving them with you, and releasing fixes to improve quality.
  • Your team is responsible for following the documentation, assigning infrastructure resources, and ensuring that in the event of an incident, you have staff familiar with the platform who can work with the Nebuly team to share context and resolve issues.

Monitoring

By default, the platform collects anonymous telemetry to help Nebuly monitor platform health and improve the product. This includes:
  • Health check information
  • Aggregate usage metrics

Upgrades

Nebuly releases new versions of the platform regularly, with improvements to performance, new features, and bug fixes. You can find the details of each release in the changelog. Nebuly recommends updating at least once per quarter. If you require support, Nebuly may ask you to upgrade to the latest version as a first step. For platform-specific upgrade instructions, see:

Hardware requirements

This section outlines the services and computing resources necessary for hosting the Platform. For high-utilization deployments (over 300k interactions per month), you may need to scale resources up significantly. Contact your account manager if you expect high traffic volumes.

Kubernetes worker node(s)

ResourceBasicHA
CPU4 vCPU4 vCPU
Memory32GB RAM32GB RAM
Instance count13

Database (PostgreSQL)

ResourceBasicHA
CPU4 vCPUs4 vCPUs
Memory16GB RAM16GB RAM
Storage size128GB128GB
Storage IOPS3,0003,000

Kubernetes Clickhouse DB node

For deployments ingesting between 100k and 300k interactions per month, a dedicated Kubernetes worker node for the ClickHouse database is required to ensure optimal performance. More details are available in the Terraform module specific to your cloud provider.
ResourceBasicHA
CPU4 vCPU4 vCPU
Memory32GB RAM32GB RAM
Instance count12
Storage typeSSDSSD
Storage IOPS4,0004,000
Storage throughput256 MB/s256 MB/s