What are Satellites?
Formal Satellites are optional specialized containers you deploy alongside Connectors to enable advanced capabilities like schema discovery and column classification. Satellites extend Formal’s core functionality without adding complexity to the Connector itself. They’re deployed in your infrastructure and communicate with Connectors and the Control Plane.Data Discovery Satellite
Automatically discovers and catalogs your database schemas, tables, columns, and relationships.Features
- Scheduled schema discovery across all resources
- Regex-based column classification from inventory path (no AI backend)
- Prompt-based column classification via AWS Bedrock
- Schema change tracking with deletion policies
Prompt Classification Backends
Prompt labels need AWS Bedrock as the LLM backend. Regex labels never need it. When prompt labels are configured, discovery samples row values from each table and sends column metadata to the classifier. The classifier returns labels based on your prompt definitions.AWS Bedrock
The Data Discovery Satellite calls AWS Bedrock directly with the default AWS credential chain. Prerequisites:- AWS credentials available to the Data Discovery Satellite container (for example, an ECS task role or EKS service account)
- Amazon Bedrock access to the Claude Haiku inference profile in your AWS account and region
FORMAL_CONTROL_PLANE_API_KEY: Satellite authentication tokenDATA_DISCOVERY_LABELING_USE_BEDROCK: Set totrueto use AWS Bedrock even when the satellite is linked to an existing AI Satellite (optional)
DATA_DISCOVERY_LABELING_USE_BEDROCK is the string true.
bedrock:InvokeModel (or equivalent) for the inference profile in your account.
Regex labels classify columns from inventory path alone and need no AI backend. Prompt labels need AWS Bedrock.
Configuration
Environment variables:FORMAL_CONTROL_PLANE_API_KEY: Satellite authentication tokenDATA_DISCOVERY_LABELING_USE_BEDROCK: Set totrueto classify prompt labels with AWS Bedrock even when the satellite is linked to an existing AI Satellite (optional)
Schema Discovery Jobs
Configure discovery schedules per resource:- Frequency: None, every 6/12/18/24 hours, or custom cron
- Deletion policy: Mark for deletion or auto-delete removed schemas
- Native user: Which credentials to use for discovery
Deployment
Satellites are Docker containers deployed in your infrastructure, similar to Connectors.Container Image Registries
You can pull the Data Discovery Satellite image from either AWS ECR or GCP Artifact Registry:On AKS, use the GAR image with
azure-gar-cred.
Set pullWithCredentials: true when using a Formal Satellite Helm chart.
For raw manifests, add formal-ecr-secret to spec.template.spec.imagePullSecrets.Prerequisites
- Create the Satellite in the Control Plane
- Copy the API token
- Deploy the container with appropriate environment variables
Recommended Deployment
- AWS ECS Fargate
- Kubernetes
- Docker
See the AWS Satellite deployment example for Terraform configuration.
Spaces and Satellites
Like Connectors and Resources, Satellites can be assigned to Spaces:- Satellite with a Space: Only communicates with Connectors and Resources in the same Space
- Satellite without a Space: Can communicate with any Connector or Resource
Managing Satellites
Creating a Satellite
1
Navigate to Satellites
Go to Satellites in the Control Plane
2
Create Satellite
Click Create Satellite
3
Configure settings
- Name: Friendly identifier
- Space: (Optional) Logical grouping
4
Copy API token
Save the token for deployment
5
Deploy container
Use the token in your deployment (ECS, Kubernetes, Docker)
Next Steps
Deploy a Satellite
Run Satellites in your infrastructure
Use in Policies
Reference Satellite data in policy rules
Configure Spaces
Segment Satellite access
View Logs
Monitor Satellite activity in logs