| eG Monitoring |
|---|
Measures reported by NutConnStatusTest The Nutanix Prism typically manages one/more Nutanix Acropolis clusters. Each node in a cluster runs a Nutanix Controller VM (CVM), which serves all of the I/O operations for the hypervisor and all VMs running on that node. A Prism service runs on every CVM with an elected Prism Leader which is responsible for handling HTTP requests. The cluster external IP will always be hosted by the Prism Leader. If the Prism Leader fails, then a new leader will be elected, which will assume the cluster external IP. This implies that the Prism service will be rendered unavailable, only if all the CVMs in a cluster fail. Such a failure can cause the dependent applications to suffer prolonged outages. Likewise, if the Prism leader takes too long to respond to HTTP requests, it will once again degrade application performance. If this is to be avoided, then the availability and responsiveness of the Prism service should be periodically checked and abnormalities should be promptly reported. This can be achieved using the NutConnStatusTest test. This test emulates an HTTP request to a Nutanix Prism, reports whether the Prism service is available or not, and if available, also reports how quickly it responds to the request. Sudden breaks in Prism availability and poor responsiveness of the Prism service can be promptly detected in the process. The test also reports the response code returned by the Prism service, so that the nature of the response - whether it is an error response or not - can be determined. The measures made by this test are as follows:
|
||||||||||||||||||||||||||||||||||||||||