For example in AWS or GCP in instance group when a vm is stopped for any reason, it is regenerated, more precisely a new instance is created to replace the one that had failed. In case there were a one-agent deploy on it will it be automatically deploy on the replacing instance or should we assign to it (instance) a new one-agent ?
Solved! Go to Solution.
If you have for example EC2 instance in AWS. After destroying it your deployment script should consider installation of one agent as well. But you have to know one thing, even if new instance will have the same ip and hostname, for DT it will be new host, so metrics will start from the beginning. Old instance will be offline.
This I’ll not affect services that are working on this host. I’m not sure how it looks for lambdas and other solutions where deployment is out of your control and it happens via PaaS token on their side.
In AWS, an EC2 is identified by instance-id, because a hostname and IP does not need to be unique overtime. Currently OneAgent is not using instance-id for host identification, but we are investigating that topic for hosts located in all cloud environments.
Technical Product Manager