Health checks

Health checks allow you to monitor the condition of your server groups, ensuring that your applications stay operational and responsive.

When creating a server group, you can define your health check protocol and monitoring intervals. The health check protocol determines how to monitor your servers.

Nebula supports the following health check protocols:

• TCP: confirms that a server is reachable over the specified port.
• HTTP: expects a specified response code after sending an HTTP request to a certain route.
• UDP_CONNECT: checks if a server can receive and respond to UDP connections.

The type of service operating on your NebCompute servers should determine the health check protocol to use. If you choose the HTTP protocol, you can optionally specify more parameters like the HTTP method, domain, port, URL path, and expected response codes.

You can also configure the monitoring interval when creating a health check. This setting determines how often the system performs the check (e.g., every 5, 10, or 30 seconds). Shorter intervals could result in quicker failure detection but may increase resource usage.

When a health check fails:

• Depending on the maximum retries specified in your configuration, the health monitor may retry a few times before updating the status of the failing instance.
• The instance is marked as unhealthy, and the status is set to stopped.
• The load balancer will stop sending traffic to it.

If the health monitor has not exceeded its maximum retries and the server starts responding successfully, the health monitor can mark it as healthy again. Once it's marked as healthy, the load balancer will resume routing traffic to it.

Need help?

If you have any technical questions or encounter any problems, get in touch with our Support team! We're here to help, and will provide support if you encounter any issues with NebCompute.