You can leverage the Azure CLI to restart the web app : https://learn.microsoft.com/en-us/cli/azure/webapp?view=azure-cli-latest#az-webapp-restart
az webapp restart --name MyWebapp --resource-group MyResourceGroup
Health Check is a feature;
The Health Check feature allows you to specify a path on your application for App Service to ping. If an instance fails to respond to the ping, the system determines it is unhealthy and removes it from the load balancer rotation. This increases your application’s average availability and resiliency.
ref: https://azure.github.io/AppService/2020/08/24/healthcheck-on-app-service.html
I'd suggest that you wouldn't want to alert administrators when an instance is replaced, as a PaaS - I would focus on the availability of your endpoint rather than the instances that comprise the service. However, as part of troubleshooting if you do experience degraded performance with the web app then i would leverage the Metrics and Logs which would allow you to view instance specific issues.