Elastic Agent doesn't resolve nodename

Elastic Observability Metrics
1

I followed to Example: Install Fleet-managed Elastic Agent on Kubernetes using Helm guide and deployed elastic agents on my k8s cluster. But agents statuses are unhealthy, and logs shows the following error

Error fetching data for metricset kubernetes.volume: error doing HTTP request to fetch 'volume' Metricset data: error making http request: Get \"https://main-cluster-dev-default-b45be-bynlj:10250/stats/summary\": lookup main-cluster-dev-default-b45be-bynlj on 10.96.0.10:53: no such host

The problem is not available in the main branch of the elastic-agent, but guide suggests to use 8.16 branch.

2

Hey @vigen, welcome to discuss :slight_smile: and sorry for the late response!

Could you try using 8.17 instead of 8.16? We may need to update the docs.

3

Hi @jsoriano, thank you for your reply!

I'll check with version 8.17 and get back with results.

4

Tested with v8.18.0, the same result :disappointed_face:

The agent tries to make a request to https://NODE_NAME:10250/stats/summary to fetch metrics. In our cluster, we're using CoreDNS with the Kubernetes plugin.
But, I couldn’t find any documentation stating that NODE_NAME will be resolved automatically. Do you know what DNS setup is assumed or required by this setup?

5

@vigen
Perhaps look at this: Kubernetes Integration - DNS Issue - Elastic Stack / Elastic Agent - Discuss the Elastic Stack not using Helm but it identifies a workaround