I have installed the Grafana agent into my Kubernetes cluster, using the Grafana Cloud Kubernetes integration as recommended in Grafana docs.
The agent is running and I can see the prometheus metrics, but I dont see any logs.
The config file of the integration only has this:
loki: configs: - name: integrations clients: - url: <client> external_labels: cluster: cloud positions: filename: /tmp/positions.yaml target_config: sync_period: 10s
I guess I need to add the scrapper configs for kubernetes.
I checked this guide, and on " Option B. Install the agent and send logs from a Kubernetes cluster", I checked at the script, but in that script the way the agent is deployed is with a DaemonSet while on the Grafana Cloud integration is a deployment.
If I use the Deployment version I guess I would only have logs for node where the agent pod is running, since it requires to mount some volumes on the host. which is not what I want.
This is the same question I have regarding node exporter. I have enabled node exporter on the Kubernetes integration, but I only see one node, which is understandable, since the agent is deployed in a single node.
I don´t understand how is this supposed to work?
Grafana Docs says that the Deployment as DaemonSet is deprecated here and should use the Kubernetes integration, which is what I am using.
Also I read somewhere in the docs that the default Kubernetes integration should keep the prometheus metrics under the free tier limit. But I have around 40k metrics now
What I am missing?