Health monitoring

Most load balancers can test the health of the application being balanced (to pre-emptively detect if a host is shut down or encounters a failure, and divert traffic away from that host).

This health test is typically done by the load balancer sending requests to specific health endpoints on the application.

MI Service Layer and MI Viewer have the following health endpoints:


Application	Heath endpoint	Response format
MI Viewer	`/mi/health`	JSON
MI Service Layer	`/mi_servicelayer/health/v3.svc/`	XML

Both endpoints return the HTTP status code 200 and will contain the text Node health OK when the MI application is able to handle incoming requests. Ansys recommend that the load balancer pings these endpoints every 10 seconds and should wait for a maximum of 30s seconds for a response.

Important: In certain scenarios, such as shortly after a new MI host has been started, the health endpoints may not report a healthy application status. This occurs because the application is still initializing and transitioning into a workable state.

Both endpoints will return an HTTP status code of 500 if the node is unable to process requests, or the response will time out if there are network issues.