Health checks are only available for servers.

​
Overview

Every time an installation is redeployed (through a Git push, configuration change, etc.), a new set of replicas will replace the old ones. To ensure zero-downtime deployments, Ryvn follows this process:

1

New Deployment

When you push an update, Ryvn creates new replicas of your installation.
2

Health Checks

Ryvn sends requests to your health check endpoint to verify the new replicas are ready.

3

Traffic Management

No traffic is routed to new replicas until they return successful responses.
4

Completion

Only after health checks pass, the old replicas are removed.

Health checks are essential for zero-downtime deployments. Every server should implement a health check endpoint so Ryvn can verify its readiness to handle traffic. Without health checks, there’s no reliable way to ensure new replicas are fully operational before routing requests to them.

​
Configuring Health Checks

To configure health checks for your installation:

1

Navigate to Environments

Go to the Environments tab in the Ryvn Dashboard
2

Select Environment

Choose the environment containing your installation
3

Select Installation

Click on the installation you want to configure
4

Configure Health Check

  1. In the Health Checks section, enter your health check endpoint path (e.g., /healthz)
  2. Click Save to apply the changes

The actions your endpoint should take to verify health depend on your installation’s details. We recommend performing operation-critical checks, such as executing a simple database query to confirm connectivity.

​
Health Check Protocol

With every health check, Ryvn sends an HTTP GET request to each replica’s health check endpoint. The response determines the replica’s health status:

CheckResponseStatus
✅ Success2xx or 3xx status code within 5 secondsHealthy
❌ FailureAny other status code or no response within 5 secondsPotentially Unhealthy

When a replica becomes potentially unhealthy, Ryvn takes different actions depending on the context:

ContextTime PeriodActions
Zero-Downtime Deployment15 consecutive minutes• Deployment is canceled
• Traffic continues flowing to existing replicas
• You are notified of the cancellation
Running Installation15 consecutive seconds• Traffic stops routing to the unhealthy replica
• Replica gets an opportunity to recover
Running Installation60 consecutive seconds• Replica is automatically restarted
• You are notified of the restart