Try it free

Kong Gateway monitoring

  • Latest Dynatrace
  • 4-min read

To enable Kong Observability in Dynatrace, you have the following options.

  • Recommended Enable OneAgent for Gateway logs, traces, and process monitoring. This should be combined with Dynatrace Prometheus scraping to monitor Kong Gateway metrics.
  • Monitor Kong using a combination of OpenTelemetry for traces and Prometheus for Kong Gateway metrics.

Kong Gateway process and logs

OneAgent automatically monitors the Kong Gateway process and logs.

Prerequisites

  • Kong Gateway version 2.8+
  • OneAgent or Dynatrace Operator is installed and available for monitoring your Kong Gateway.

The required installation depends on your application:

If your application is runningSee the instruction for
on a virtual machine or bare-metalOneAgent
as workload in Kubernetes or OpenShiftDynatrace Operator

Application traces

In addition to process and logs, OneAgent also provides Kong Gateway application traces. See Manual runtime instrumentation for NGINX.

Step 1 Configure Kong Gateway

Prerequisites

  • Kong Gateway version 3.8+

Kong Gateway requires the configuration of two settings.

  • tracing_instrumentations = all
  • tracing_sampling_rate = 1.0

For further details and option, see the Kong Gateway documentation.

Step 2 Configure OpenTelemetry plugin

  1. Evaluate support for logging and tracing according to the OpenTelemetry plugin version installed in your environment.

  2. Send the following POST request (example assumes Kong Gateway 3.8+) by replacing {HOST}, {PLUGIN-INSTANCE_NAME} and {OPENTELEMETRY_COLLECTOR} with proper values:

curl -X POST http://{HOST}:8001/plugins \
 -H 'Content-Type: application/json' \
 -d '{
   "name": "opentelemetry",
   "instance_name": "{PLUGIN-INSTANCE_NAME}",
   "config": {
     "traces_endpoint": "http://{OPENTELEMETRY_COLLECTOR}:4318/v1/traces",
     "logs_endpoint": "http://{OPENTELEMETRY_COLLECTOR}:4318/v1/logs",
     "resource_attributes": {
       "service.name": "kong-dev"
      }
    }
  }'

Step 3 Configure OpenTelemetry Collector

Configure your OpenTelemetry Collector to send data to your Dynatrace environment. The example below shows how to export traces and logs.

receivers:
  otlp:
    protocols:
      http:
        endpoint: 0.0.0.0:4318


exporters:
  otlp_http:
    endpoint: "${env:DT_BASEURL}/api/v2/otlp"
    headers:
      "Authorization": "Api-Token ${env:DT_API_TOKEN}"


service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: []
      exporters: [otlp_http]
    logs:
      receivers: [otlp]
      processors: []
      exporters: [otlp_http]

Step 3 Export Application Span Metrics

To include span metrics for application traces, configure the collector exporters section of the OpenTelemetry Collector configuration.

connectors:
  spanmetrics:
    dimensions:
      - name: http.method
        default: GET
      - name: http.status_code
      - name: http.route
    exclude_dimensions:
      - status.code
    metrics_flush_interval: 15s
    histogram:
      disable: false


service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: []
      exporters: [spanmetrics]
    metrics:
      receivers: [spanmetrics]
      processors: []
      exporters: [otlp_http]

Metrics

Kong’s Prometheus plugin is a convenient way to collect Kong Gateway metrics. Dynatrace can collect these metrics directly from the Gateway produced by the Kong plugin. The default port and endpoint is 8001/metrics.

For more information and a list of available metrics, see the Kong Prometheus plugin documentation.

Step 1 Enable Kong Prometheus plugin

Basic configuration

To enable basic configuration of the Kong Prometheus plugin send a POST request replacing {HOST} with the host name value.

curl -s -X POST http://{HOST}:8001/plugins \
-H 'Content-Type: application/json' \
-d '{
  "name": "prometheus"
}'

Additional plugin metrics

To enable additional metrics produced by the Kong Gateway Prometheus plugin, send a POST request replacing {HOST} and {PLUGIN-INSTANCE_NAME} with proper values:

curl -s -X POST http://{HOST}:8001/plugins \
-H 'Content-Type: application/json' \
-d '{
  "name": "prometheus",
  "instance_name": "{PLUGIN-INSTANCE_NAME}",
  "config": {
  "per_consumer": true,
  "status_code_metrics": true,
  "latency_metrics": true,
  "bandwidth_metrics": true,
  "upstream_health_metrics": true
 }
}'

To check available Kong metrics, query the /metrics endpoint:

curl -i http://{HOST}:8001/metrics

Step 2 Collect Prometheus metrics

After configuring Kong Gateway's Prometheus plugin, metrics can be collected using the Dynatrace ActiveGate (recommended) or the OpenTelemetry Collector.

Scrape metrics using ActiveGate

In Kubernetes, Dynatrace supports scraping of Prometheus endpoints using special annotations.

To learn how to collect Prometheus metrics in Kubernetes, see Monitor Prometheus metrics.

Scrape metrics using OpenTelemetry Collector

You can also use the OpenTelemetry Collector’s Prometheus receiver to collect metrics from the Kong Gateway. To learn how to scrape Prometheus data using an OpenTelemetry collector, see Scrape Prometheus metrics with the OTel Collector.

If you're running on Kubernetes, you can enrich traces, metrics, and logs using the Collector's Kubernetes attribute processor. This allows Dynatrace to map the telemetry data to the correct toplogy. To learn how to enable enrichment in the OpenTelemetry Collector, see Enrich OTLP requests with Kubernetes data.