setup-runtime.md

Setup Runtime Visibility

Now that we've got visibility into the early stages of the development lifecycle, lets move onto runtime visibility! The Mountain Passes stack in this repository is designed to be deployed to a Kubernetes cluster using ArgoCD. In order to run it, you'll need an appropriate Kubernetes setup.

In particular, you will need the Datadog Operator (to collect logs, traces, and metrics) and ArgoCD (to deploy the stack) installed and running. The easiest way to get started is using the Datadog operator and the ArgoCD getting started guided. Given an existing cluster and familiarity with Kubernetes these steps shouldn’t take too long. Your operator configuration will need at least the following flags configured in order to fully support the app:

apiVersion: datadoghq.com/v2alpha1
kind: DatadogAgent
metadata:
  name: datadog
spec:
  override:
    nodeAgent:
      env:
        # Ensures that the agent will accept logs over OTLP, necessary
        # for the pass-image-api
        - name: DD_OTLP_CONFIG_LOGS_ENABLED
          value: "true"
  global:
    # Make sure to set the site correctly if you are not using Datadog US!
    site: datadoghq.com

    # Set your credentials up normally. For example ... 
    credentials:
      apiSecret:
        secretName: datadog-secret
        keyName: api-key
      appSecret:
        secretName: datadog-secret
        keyName: app-key
  features:
    # Ensure the gRPC receive is enabled for OTLP. The pass-image-api needs this
    otlp:
        receiver:
          protocols:
            grpc:
              enabled: true
              hostPortConfig:
                enabled: true

    # Ensure APM is enabled, so that we can collect traces
    apm:
      enabled: true

    # Log collection is needed for pass-api and pass-summary-api to export their logs
    logCollection:
      enabled: true
      containerCollectAll: true

    # These are not directly related but are fairly typical
    liveProcessCollection:
      enabled: true
    processDiscovery:
      enabled: true
    kubeStateMetricsCore:
      enabled: true 
    clusterChecks:
      enabled: true
    eventCollection:
      collectKubernetesEvents: true

Tip

If you'd like to try this out on your laptop, minikube is a great tool to quickly spin up a local cluster, and will work well with everything needed for this demo.

Once you've got Datadog and ArgoCD in place, you can move on to deploying the app!