grafana / intro-to-mltp

Introduction to Metrics, Logs, Traces and Profiles session companion code.
GNU Affero General Public License v3.0
327 stars 71 forks source link

Example for Kubernetes #92

Open y6vmeq opened 3 months ago

y6vmeq commented 3 months ago

The example perfectly shows the capabilities of the stack in action. However, there are problems in reproducing the same results in kubernetes cluster. It would be great if there were pre-configured Helm charts or yaml manifests To reproduce all the same features in Kubernetes.

hedss commented 3 months ago

Hi @y6vmeq Apologies for the delay in response to this issue.

We've actually discussed at length whether or not to use this repo as a forum for also deploying to Kubernetes (and, indeed, there are manifests for deploying the application itself to Kubernetes).

However, there is already good documentation for doing this which point to the relevant Helm charts, such as for Tempo here.

We could potentially produce a Helm chart that deploys all of the components in this repository, but the worry is due to the number of moving parts this would be a fairly heavy maintanance workload (though saying that, we could potentially use some GH actions to automatically generate dependency updates). We would, though, have to keep updates in-line with the Docker Compose manifests, and some of the versioning we have is due to changes in functionality and therefore is occasionally staggered with mainline releases.

I think we'll reopen some discussion internally on this and make a yes/no decision in the (fairly) near future.

Thank you for reaching out!

eyalsolomon1 commented 2 months ago

@hedss +1 on @y6vmeq request, not sure a single chart is needed, rather a 'gist' style of using existing charts and "manually" adding each component with it's relevant configs