Kong Gateway monitoring

With Dynatrace you can get observability for Kong Gateway and its managed APIs. Learn how to trace your Kong Gateway and collect metrics using the Kong Prometheus extension.

Prerequisites

Your Kong Gateway version is on 2.1+
OneAgent or Dynatrace Operator is installed and available for monitoring your Kong Gateway.

The required installation depends on your application:

If your application is running See the instruction for
on a virtual machine or bare-metal OneAgent
as workload in Kubernetes or OpenShift Dynatrace Operator

If your application is running	See the instruction for
on a virtual machine or bare-metal	OneAgent
as workload in Kubernetes or OpenShift	Dynatrace Operator

Traces

To get started with tracing of Kong applications, see Manual runtime instrumentation of NGINX.

Metrics

Dynatrace can collect metrics using the Kong Prometheus plugin. Thereby, the following configuration is assumed:

Setting	Details
Kong Services Endpoint	`/services`
Kong Routes Endpoint	`/routes`
Kong Plugins Endpoint	`/plugins`
Prometheus Metrics Endpoint	`/metrics`
Kong Proxy HTTP Port	`8000`
Kong Proxy HTTPS Port	`8443`
Kong Admin HTTP Port	`8001`
Kong Admin HTTPS Port	`8444`

Make sure to adapt the installation instructions to your actual configuration.

Enable Prometheus plugin on Kong

To enable the Prometheus plugin on Kong, send the following POST request to the /plugins endpoint:

shell
curl -X POST http://{HOST}:8001/plugins --data "name=prometheus"

Replace {HOST} with the your actual value.

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

shell
curl -i http://localhost:8001/metrics

Collect Prometheus metrics

The Prometheus plugin is the most convenient way to collect Kong metrics. The plugin is under continuous development, so more and more new metrics are coming.

Dynatrace currently provides three approaches for collecting Prometheus metrics, depending on the environment.

KubernetesUsing ExtensionAmazon Managed

In Kubernetes, Dynatrace supports scraping of any HTTP(s) endpoint offering metrics in OpenMetrics format (for example, Prometheus exporters). Using Dynatrace-specific annotations, you can specify which pods or services to scrape.

Learn how to collect Prometheus metrics in Kubernetes.

Dynatrace provides a scalable means to ingest Prometheus metrics directly from the source, without Kubernetes. This works fully automatically and works best if OneAgent is installed on the box where the Prometheus metrics originate, but it can also be done in a fully agentless manner when OneAgent can't be installed on the box.

Learn how to collect Prometheus metrics without Kubernetes using Extensions 2.0 Prometheus data source.

Check our Extensions 2.0 tutorials.

Extension example based on Prometheus

yaml
name: custom:com.dynatrace.extension.micrometer-quarkus
version: 1.0.0
minDynatraceVersion: "1.247"
author:
  name: Dynatrace

#dashboards:
#  - path: "dashboards/dashboard_exporter.json"

#alerts:
#  - path: "alerts/alert_socket_usage.json"

prometheus:
  - group: quarkus metrics
    interval:
      minutes: 1
    featureSet: all
    dimensions:
      - key: quarkus
        value: const:quarkus
    subgroups:
      # global counters
      - subgroup: quarkus global counter
        dimensions:
          - key: global_counters
            value: const:global_counters
        metrics:
          # HELP process_uptime_seconds The uptime of the Java virtual machine
          # TYPE process_uptime_seconds gauge
          - key: com.dynatrace.process.global.uptime.seconds
            value: metric:process_uptime_seconds
            type: gauge
            featureSet: global

          # HELP process_cpu_usage The "recent cpu usage" for the Java Virtual Machine process
          # TYPE process_cpu_usage gauge
          - key: com.dynatrace.process.global.cpu.usage
            value: metric:process_cpu_usage
            type: gauge
            featureSet: global

          # HELP system_cpu_usage The "recent cpu usage" of the system the application is running in
          # TYPE system_cpu_usage gauge
          - key: com.dynatrace.system.global.cpu.usage
            value: metric:system_cpu_usage
            type: gauge
            featureSet: global

          # HELP jvm_classes_unloaded_classes_total The total number of classes unloaded since the Java virtual machine has started execution
          # TYPE jvm_classes_unloaded_classes_total counter
          - key: com.dynatrace.jvm.classes.global.uploaded.total
            value: metric:jvm_classes_unloaded_classes_total
            type: count
            featureSet: global

          # HELP jvm_info_total JVM version info
          # TYPE jvm_info_total counter
          - key: com.dynatrace.jvm.global.info.total
            value: metric:jvm_info_total
            type: count
            featureSet: global

          # HELP http_server_connections_seconds_max
          # TYPE http_server_connections_seconds_max gauge
          - key: com.dynatrace.http.server.connections.seconds.global.max
            value: metric:http_server_connections_seconds_max
            type: gauge
            featureSet: global

          # HELP http_server_connections_seconds
          # TYPE http_server_connections_seconds summary
          - key: com.dynatrace.http.server.connections.seconds.active.global.count
            value: metric:http_server_connections_seconds_active_count
            type: count
            featureSet: global
          - key: com.dynatrace.http.server.connections.seconds.active.global.duration.summary
            value: metric:http_server_connections_seconds_duration_sum
            type: gauge
            featureSet: global

          # HELP process_files_max_files The maximum file descriptor count
          # TYPE process_files_max_files gauge
          - key: com.dynatrace.process.files.global.max
            value: metric:process_files_max_files
            type: gauge
            featureSet: global

          # HELP http_server_bytes_written_max
          # TYPE http_server_bytes_written_max gauge
          - key: com.dynatrace.http.server.bytes.wrriten.global.max
            value: metric:http_server_bytes_written_max
            type: gauge
            featureSet: global

          # HELP http_server_bytes_written
          # TYPE http_server_bytes_written summary
          - key: com.dynatrace.http.server.bytes.written.global.count
            value: metric:http_server_bytes_written_count
            type: count
            featureSet: global
          - key: com.dynatrace.http.server.bytes.written.global.summary
            value: metric:http_server_bytes_written_sum
            type: gauge
            featureSet: global

          # HELP system_load_average_1m The sum of the number of runnable entities queued to available processors and the number of runnable entities running on the available processors averaged over a period of time
          # TYPE system_load_average_1m gauge
          - key: com.dynatrace.system.load.average.global.lm
            value: metric:system_load_average_1m
            type: gauge
            featureSet: global

          # HELP jvm_gc_overhead_percent An approximation of the percent of CPU time used by GC activities over the last lookback period or since monitoring began, whichever is shorter, in the range [0..1]
          # TYPE jvm_gc_overhead_percent gauge
          - key: com.dynatrace.jvm.gc.overhead.global.percent
            value: metric:jvm_gc_overhead_percent
            type: gauge
            featureSet: global

          # HELP jvm_threads_daemon_threads The current number of live daemon threads
          # TYPE jvm_threads_daemon_threads gauge
          - key: com.dynatrace.jvm.threads.daemon.global.threads
            value: metric:jvm_threads_daemon_threads
            type: gauge
            featureSet: global

          # HELP jvm_threads_live_threads The current number of live threads including both daemon and non-daemon threads
          # TYPE jvm_threads_live_threads gauge
          - key: com.dynatrace.jvm.threads.live.global.threads
            value: metric:jvm_threads_live_threads
            type: gauge
            featureSet: global

          # HELP http_server_requests_seconds
          # TYPE http_server_requests_seconds summary
          - key: com.dynatrace.http.server.bytes.written.global.count
            value: metric:http_server_requests_seconds_count
            type: count
            featureSet: global
          - key: com.dynatrace.http.server.bytes.written.global.summary
            value: metric:http_server_requests_seconds_sum
            type: gauge
            featureSet: global

          # HELP http_server_requests_seconds_max
          # TYPE http_server_requests_seconds_max gauge
          - key: com.dynatrace.http.server.requests.seconds.max
            value: metric:http_server_requests_seconds_max
            type: gauge
            featureSet: global

          # HELP process_start_time_seconds Start time of the process since unix epoch.
          # TYPE process_start_time_seconds gauge
          - key: com.dynatrace.process.start.time.global.seconds
            value: metric:process_start_time_seconds
            type: gauge
            featureSet: global

          # HELP jvm_classes_loaded_classes The number of classes that are currently loaded in the Java virtual machine
          # TYPE jvm_classes_loaded_classes gauge
          - key: com.dynatrace.jvm.classes.loaded.global.max
            value: metric:jvm_classes_loaded_classes
            type: gauge
            featureSet: global

          # HELP jvm_threads_peak_threads The peak live thread count since the Java virtual machine started or peak was reset
          # TYPE jvm_threads_peak_threads gauge
          - key: com.dynatrace.jvm.threads.peak.global.threads
            value: metric:jvm_threads_peak_threads
            type: gauge
            featureSet: global

          # HELP system_cpu_count The number of processors available to the Java virtual machine
          # TYPE system_cpu_count gauge
          - key: com.dynatrace.system.cpu.global.counter
            value: metric:system_cpu_count
            type: gauge
            featureSet: global

          # HELP process_files_open_files The open file descriptor count
          # TYPE process_files_open_files gauge
          - key: com.dynatrace.process.files.open.global.files
            value: metric:process_files_open_files
            type: gauge
            featureSet: global

Also, keep watching our Software intelligence hub, where we frequently publish new extensions.

Amazon provides a managed service for Prometheus, and Dynatrace provides the means to collect metrics directly from it if you need it. Typically, however, it would be more cost-effective to use one of the other methods described here and avoid the additional cost of the Amazon service.

Learn how to collect Prometheus metrics from Amazon Managed Service for Prometheus.

Kong metrics

Metric key	Description	Type
`kong_bandwidth`	Total bandwidth in bytes consumed per service/route in Kong	counter
`kong_datastore_reachable`	Datastore reachable from Kong; 0 is unreachable	gauge
`kong_http_status`	HTTP status codes per service/route in Kong	counter
`kong_memory_lua_shared_dict_bytes`	Allocated slabs in bytes in a shared_dict	gauge
`kong_memory_lua_shared_dict_total_bytes`	Total capacity in bytes of a shared_dict	gauge
`kong_memory_workers_lua_vms_bytes`	Allocated bytes in worker Lua VM	gauge
`kong_nginx_http_current_connections`	Number of HTTP connections	gauge
`kong_nginx_metric_errors_total`	Number of nginx-lua-prometheus errors	counter
`kong_nginx_timers`	Number of NGINX timers	gauge