Is there a way to add multiple fleet hosts endpoint in the fleet and create a separation based on the agent policies?
Our current Elastic deployment hosted on Elastic cloud has default fleet hosts and the agents are enrolled to the fleet and working as expected.
We have a new requirement to install elastic agent on Google environment and setup private communications between GCP and Elastic cloud using private service connect - GCP Private Service Connect traffic filters | Elasticsearch Service Documentation | Elastic
The problem with this setup is that, we need to add a new fleet host to the fleet - https://<Fleet cluster ID/deployment alias>.fleet.:443. When we add the fleet host, the existing agents enrolled with the default fleet host goes into unhealthy state because the agents try to connect to the new PSC host URL. The agents go into a unhealthy state because the DNS for the fleet host is hosted in the GCP private zone and existing agents are unaware of this domain.
Here is the error message from the existing Elastic agent
[elastic_agent][error] failed to dispatch actions, error: fail to communicate with updated API client hosts: Get "https://<Fleet cluster ID>.fleet.<private zone DNS name>:443/api/status?": lookup <Fleet cluster ID>.fleet.<private zone DNS name: Temporary failure in name resolution