One of the most used oberservability stacks in Kubernetes must be the Kube-Prometheus-Stack. Bringing Prometheus, Grafana, Alertmanager to the table, Promtail and Loki completing the stack on the logging side.

One problem I always had with this is its resource consumption. It might not be as huge as one might expect it to be, but still if you are running a really low power homelab on k3s this might be important to you.

Are there any alternatives that consume way less resources? Even if they might not be as all powerful as Grafana?

Yes there is!

The Dotnet Aspire Dashboard has originally been developed to be a leightweight tool to aid local development. And bring the local development environment closer to the desired cloud setups.

So my goal was combining the Dotnet Aspire Dashboard with the OpenTelemetry Collector to see how much functionality this might be able to provide.

So I set out on creating this Helm Chart, here is whats currently possible using this.

Log collection by application

The OpenTelemetry Collector is able to collect the logs for all Pods running in your cluster and forwarding them to the Dotnet Aspire Dashboard.

Host metrics

The OpenTelemetry Collector is able to scrape host metrics and send them to the Dotnet Aspire Dashboard. At the moment Pod metrics do not yet work, except you configure your application to send send your metrics to the Dashboard using OpenTelemetry.

For the future

Pod metrics

Would it be possible to also scrape Pod metrics without the need to change the pod configuration?