Connector health checks

Monitor Baton connectors running in service mode with HTTP health check endpoints.

Baton connectors running in service mode can expose HTTP health check endpoints for container orchestrators (ECS Fargate, Kubernetes, etc.) to monitor connector health.

When C1 encounters a 503 from a connector during provisioning, it treats this as a transient error and retries automatically (up to 5 attempts). Errors indicating a configuration problem — such as 401, 403, or 404 — are not retried. If provisioning still fails after retries, the task falls back to a manual path and appears in the Assigned to me area for follow-up. Learn more about automatic account provisioning.

Enabling health checks

Health checks are disabled by default. Enable them using environment variables or CLI flags:

Environment variable	CLI flag	Default	Description
BATON_HEALTH_CHECK	--health-check	false	Enable the health check server
BATON_HEALTH_CHECK_PORT	--health-check-port	8081	Port for the health check server
BATON_HEALTH_CHECK_BIND_ADDRESS	--health-check-bind-address	127.0.0.1	Bind address for the health check server

Health check endpoints

Endpoint	Description	Success (HTTP 200)	Failure (HTTP 503)
/health	Full health check - calls the connector’s Validate() method	Connector is healthy	Connector validation failed
/ready	Readiness check - verifies connector client is available	Connector is ready to serve requests	Connector client not available
/live	Liveness check - verifies process is running	Always returns success	N/A

The health-check command

Each connector includes a health-check subcommand for querying the health check server. This is designed for container exec probes, eliminating the need to bundle curl or other HTTP clients in container images.

Usage

# Check health (default endpoint)
baton-myconnector health-check

# Check readiness
baton-myconnector health-check --endpoint=ready

# Check liveness
baton-myconnector health-check --endpoint=live

# Check with custom port
baton-myconnector health-check --health-check-port=9090

Flags

Flag	Default	Description
--endpoint	health	Endpoint to check: health, ready, or live
--timeout	5	Request timeout in seconds
--health-check-port	8081	Port to connect to
--health-check-bind-address	127.0.0.1	Address to connect to

Exit codes

Scenario	Exit code
HTTP 200 response	0 (success)
HTTP non-200 response	1 (failure)
Connection failed	1 (failure)
Timeout	1 (failure)

Container orchestrator examples

AWS ECS Fargate

{
  "healthCheck": {
    "command": ["CMD", "/baton-myconnector", "health-check"],
    "interval": 30,
    "timeout": 10,
    "retries": 3,
    "startPeriod": 60
  }
}

Kubernetes

livenessProbe:
  exec:
    command: ["/baton-myconnector", "health-check", "--endpoint=live"]
  initialDelaySeconds: 10
  periodSeconds: 5

readinessProbe:
  exec:
    command: ["/baton-myconnector", "health-check", "--endpoint=ready"]
  initialDelaySeconds: 5
  periodSeconds: 3