Enabling Datadog integration for ingress-nginx Helm chart

This is a guide on how to enable the Datadog nginx and nginx_ingress_controller integrations when using the ingress-nginx Helm chart.

Available Integrations in Datadog

In Datadog, there are two NGINX-related integrations available:

• nginx: Metrics reported by NGINX status page. The required ngx_http_stub_status_module is enabled by default in the ingress-nginx Helm chart. Access to the status page is restricted, and we need to whitelist the Datadog agent's IP. The list of metrics can be found in the nginx documentation.
• nginx_ingress_controller: Metrics reported by the NGINX Ingress controller. It exposes metrics in the Prometheus format. The list of metrics can be found in a GitHub PR.

In this guide, we will enable both integrations.

Enabling metrics endpoint and status page

Metrics for the NGINX Ingress controller are disabled by default. We have to enable the Prometheus metrics endpoint:

controller:
  metrics:
    enabled: true

For NGINX, we have to whitelist the IP of the Datadog agent to allow access to the status page. In the following example, I've simply whitelisted all IPs in the Kubernetes VPC so that any Pod can access it:

controller:
  config:
    nginx-status-ipv4-whitelist: "10.1.0.0/16"

Agent check configuration

The Datadog agent looks for a ad.datadoghq.com/<CONTAINER_IDENTIFIER>.checks annotation on Pods to check for integration configurations.

The CONTAINER_IDENTIFIER is the name of the container that the integration applies to. By default, the container name of the NGINX Ingress controller is controller so our annotation is named ad.datadoghq.com/controller.checks.

For the Datadog nginx integration, we have to set the nginx_status_url. The path is always /nginx_status.

For the nginx_ingress_controller integration, we have to set the URL of the Prometheus metrics endpoint. By default, they are exposed on port 10254. The path is always /metrics.

controller:
  podAnnotations:
    ad.datadoghq.com/controller.checks: |
      {
        "nginx": {
          "instances": [{"nginx_status_url": "http://%%host%%/nginx_status"}]
        },
        "nginx_ingress_controller" : {
          "instances": [{"prometheus_url": "http://%%host%%:10254/metrics"}]
        }
      }

The %%host%% part is a template variable that will be replaced automatically with the agent's host service IP. This will only work if the agent is running on every host.

Check if everything works

We can check if the metrics collection works by running the agent status command in an agent container. If everything is working, it will report a successful execution date:

Last Execution Date : 2022-07-27 03:43:24 UTC (1658893404000)
Last Successful Execution Date : 2022-07-27 03:43:24 UTC (1658893404000)

Finally, we should enable the integrations in the Datadog Web UI. Go to "Integrations", select nginx/nginx-ingress-controller and click on "Enable". Doing this will also create 3 default dashboards.