Skip to main content
HoneyHive offers self-hosted deployments for organizations with strict privacy, security, or compliance requirements. Both the Control Plane and Data Plane are deployed in your environment, giving you full control over all infrastructure and data.
Self-hosted deployments are managed through our Enterprise team. Contact sales@honeyhive.ai to get started.

Deployment models

We support two deployment models:
ModelDescription
BYOC (Bring Your Own Cloud)We deploy HoneyHive’s Kubernetes cluster in a dedicated cloud account you provision (AWS, GCP, or Azure).
BYOVPC (Bring Your Own VPC)We deploy HoneyHive’s Kubernetes cluster within your existing VPC, with VPC peering to your infrastructure.
Both models give you full network isolation and control over data residency.

Supported platforms

Cloud ProviderKubernetesDeployment
AWSEKSTerraform + Helm
GCPGKETerraform + Helm
AzureAKSTerraform + Helm
On-premiseRKE, OpenShiftHelm
Need a different platform? Let us know.

What’s included

During setup, we work with your security and infrastructure teams to:
  • Review network architecture and data flow against your policies
  • Configure Terraform modules and Helm charts to meet your requirements
  • Deploy the cluster and validate the installation
  • Set up VPC peering (for BYOVPC) or PrivateLink
  • Integrate SAML/SSO with your identity provider
See Platform Architecture for details on how these components work together. For more on our security posture and compliance, see Security.

SDK Configuration

In federated deployments, the Control Plane and Data Plane may run at different URLs. The SDK client accepts separate CP and DP base URLs, though not every SDK surface routes through the CP URL yet - currently it is used on the events read path. For span ingestion and most API calls, HH_API_URL is the primary endpoint.

Environment Variables

# Data Plane URL - used for span/event ingestion (required)
export HH_API_URL="https://dp.your-company.honeyhive.ai"

# Control Plane URL (optional) - accepts a separate CP endpoint,
# though not all SDK surfaces route through it yet.
# Defaults to HH_API_URL if not set.
export HH_CP_API_URL="https://cp.your-company.honeyhive.ai"

export HH_API_KEY="your-api-key"

SDK Parameters

You can also pass URLs directly when initializing the SDK client:
from honeyhive import HoneyHive
import os

client = HoneyHive(
    api_key=os.getenv("HH_API_KEY"),
    base_url="https://dp.your-company.honeyhive.ai",
    cp_base_url="https://cp.your-company.honeyhive.ai",
)
For the tracer, use server_url:
from honeyhive import HoneyHiveTracer
import os

tracer = HoneyHiveTracer.init(
    api_key=os.getenv("HH_API_KEY"),
    project="my-project",
    server_url="https://dp.your-company.honeyhive.ai",
)
The client resolves HH_CP_API_URL (falling back to HH_API_URL) for CP-bound requests, but not every SDK surface uses it yet. For now, most operations - including tracer session creation - still go through HH_API_URL. Set HH_CP_API_URL if your deployment splits CP and DP hosts so the SDK is ready when full routing support lands.

Resolution Order

The SDK resolves URLs in this order (first match wins):
PlaneResolution order
Data Planebase_url parameter → HH_API_URL env var → https://api.honeyhive.ai
Control Planecp_base_url parameter → HH_CP_API_URL env var → HH_API_URL env var → https://api.honeyhive.ai
For cloud-hosted HoneyHive (default), you don’t need to set any URL - both planes share https://api.honeyhive.ai. Separate URLs are only needed for self-hosted federated deployments where the CP and DP run on different hosts.

Maintenance and Support

Self-hosted deployments include:
  • Dedicated implementation - Setup and deployment led by our forward-deployed engineering team
  • Guided upgrade process with versioned changelogs - Versioned releases with migration guides and patch diffs provided for your team to review and apply
  • Custom support packages - Tailored support availability and incident response policies to meet your organization’s requirements

Get started

Contact Enterprise Sales

Reach out to discuss your deployment requirements.