Add Simple kubernetes setup

[dineshg13]

Fixes a part of #153
Adds kubernetes files to test webstore in a k8s cluster.

Changes

1. Adds a new directory k8s_local with the setup necessary to run the demo webstore on a k8s cluster.
2. Adds changes to Makefile to make easy to setup the services in k8s

Cons of the approach:

1. We now a different config files for running services via docker-compose and k8s but that is inevitable as some of config are different are different in two environments.

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: dineshg13 / name: Dinesh Gurumurthy (b4886a4, f2473cd)

[fatsheep9146]

I think the image tag version should be a variable, and keep up to date.

[dineshg13]

Made this an env var that we set in Makefile. Let me know if there is a better way to achieve same.

[wph95]

Communicated with the author via slack. It was agreed that the pr will be continued after the otel demo helm chart is completed.

[saurabhdes]

I tested this on my Macbook Pro after creating a local cluster via kind. I see deployments getting created (e.g. for PaymentService, frontend etc. but no pods. I ran the command make k8s_local_apply . Not sure if i am missing something. When i do a describe on deployment e.g. kubectl describe deployment paymentservice, i noticed the image name seems incorrect otel/demo:-paymentservice and export DOCKER_IMAGE_VERSION=v0.2.0-alpha doesn't seem to be applied to the yaml file.

[dineshg13]

@saurabhdes thanks for testing. I made a fix to makes sure env var is set properly in Makefile . Could you retest this ?

[saurabhdes]

thanks @dineshg13 for the fix. I am still having same issues with my kind cluster (could be my machine issue). But i was able to deploy successfully on AKS cluster.

Not all services were there as i am running into a separate memory issue for the otel daemon set to run. I will try increasing memory size of AKS nodes.

[fatsheep9146]

Not all services were there as i am running into a separate memory issue for the otel daemon set to run. I will try increasing memory size of AKS nodes.

Maybe consider only using limit for resources not request to avoid this kind of problem? @dineshg13

[saurabhdes]

I removed this

resources:
            limits:
              cpu: 1
              memory: 2Gi
            requests:
              cpu: 100m
              memory: 400Mi

"resources" section from k8s_local/otel-agent-daemonset.yaml and after that the all the otel collector daemonset started. Then i saw more accurate flowmap in Jaeger.

[fatsheep9146]

I removed this

resources:
            limits:
              cpu: 1
              memory: 2Gi
            requests:
              cpu: 100m
              memory: 400Mi

"resources" section from k8s_local/otel-agent-daemonset.yaml and after that the all the otel collector daemonset started. Then i saw more accurate flowmap in Jaeger.

Although this works, but I think it's the yaml with resources section is still important, since the resources spec is also important observation information. I think maybe a flag to switch between with-resource mode and without-resource mode is better. @dineshg13 @wph95

[JaredTan95]

following #156, the package is now public, we could use ghcr.io/open-telemetry/demo now.

[dineshg13]

Closing this in favor of open-telemetry/opentelemetry-helm-charts#280