Continuing the discussion from Agent stuck on updating but still send index:
We are also seeing this. We have Fleet deployed in ECK. We have 2 fleet servers listed in Kibana, one url is the k8s internal Cluster DNS FQDN for fleet server. This is being used by agents running within the k8s cluster. The other is the External IP for agent for agents running on other hosts. The config is below.
The agents internal to the cluster that report directly to Elasticsearch are healthy. The external agents are sending data through Logstash never change from "Upgrading" to healthy after install. The data from the integrations is being indexed and the logs on the agent show that its pulling the policy updates as does Kibana show that the agent is checking in. It appears the only difference between the agents in the use of Logstash.