Add a walkthrough for running an otel-collector inside Kubernetes

[XZymo]

Suggested improvements for page: Kubernetes

There doesn't seem to be a walkthrough for running an otel-collector inside Kubernetes for receiving traces from external clients, only for monitoring a Kubernetes cluster.

[tiffany76]

Hi @XZymo. Thanks for raising this issue. Before I escalate it, I wonder if you have reviewed the OpenTelemetry Kubernetes Demo. As mentioned in the Getting Started page, the demo is where we recommend you start learning how to implement OpenTelemetry in a Kubernetes deployment. Let us know if the demo page doesn't answer your questions.

[XZymo]

Hello! I did test out the demo and was able to set it up using the demo helm chart, however I found that the document doesn't goes go very far in details about the way the service is deployed and why the service was configured that way. The document seems to go more in detail about how to configure an ingress / service type / browser telemetry for the demo.

My point is there is no walkthrough for the opentelemetry-collector's helm chart specifically and how to configure it for my own purposes. In fact, the collector's README only really delves into configuring the collector for Kubernetes clusters and not running the collector for external services/clients.

[tiffany76]

@open-telemetry/helm-approvers, @open-telemetry/collector-approvers: do we have any examples or documentation in the code repos that could help @XZymo?

[jpkrohling]

Typically, people would deploy the Collector either via the Operator or via the Helm Charts on Kubernetes. Then, from there, they can expose the Collector via regular Kubernetes mechanisms (services + ingresses). The only thing that is specific to the Collector would be to secure it, to which I think we have instructions. If not, this is a blog post I know was up to date as of a couple of months ago: https://medium.com/opentelemetry/securing-your-opentelemetry-collector-1a4f9fa5bd6f

[svrnm]

following up on this issue, can we improve our documentation around that somehow or can this be closed?

[meswapnilk]

Hello! I did test out the demo and was able to set it up using the demo helm chart, however I found that the document doesn't goes go very far in details about the way the service is deployed and why the service was configured that way. The document seems to go more in detail about how to configure an ingress / service type / browser telemetry for the demo.

My point is there is no walkthrough for the opentelemetry-collector's helm chart specifically and how to configure it for my own purposes. In fact, the collector's README only really delves into configuring the collector for Kubernetes clusters and not running the collector for external services/clients.

I don't think there is anything beyond configuring the service/ingress for the collector, the same can be used to send the metrics data from any kubernetes/non-kubernetes services/clients. How to use the configuration is part of your instrumentation using either language SDKs or Zero-code Instrumentation

[tiffany76]

Closing this for now. If you still have suggestions for improvement to the documentation, please re-open or raise a new issue. Thanks!