Skip to content

Kong

Collect Kong metrics, logs, and tracing information

Metrics

1. Kong Configuration

To enable the reporting of Kong metrics, the Prometheus plugin must be activated. Here, it is recommended to configure this globally. You can refer to the official Kong documentation.

Configure via Kong Manager

Log in to Kong Manager, click on the Plugins menu, and add a new Prometheus.

Configure via Kong API

curl -X POST http://localhost:8001/plugins \
    --header "accept: application/json" \
    --header "Content-Type: application/json" \
    --data '
    {
  "name": "prometheus",
  "config": {
    "per_consumer": true,
    "status_code_metrics" : true,
    "bandwidth_metrics": true
  }
}'

2. DataKit Configuration

  • Navigate to the conf.d/samples directory under the DataKit installation directory, copy prom.conf.sample and rename it to kong.conf

cp prom.conf.sample kong.conf

  • Adjust kong.conf

Only modify the following two parameters; leave the others unchanged:

[[inputs.prom]]
  ## Exporter URLs.
  urls = ["http://127.0.0.1:8001/metrics"]
  keep_exist_metric_name = true
  • Restart DataKit

Execute the following command:

datakit service -R

3. Kong Metric Sets

Metric Name Unit Description
kong_bandwidth_bytes bytes Bandwidth usage of Kong, unit in bytes
kong_datastore_reachable - Status metric indicating whether the data store is reachable
kong_http_requests_total count Total number of HTTP requests
kong_kong_latency_ms_bucket ms Histogram bucket for Kong's own latency, unit in milliseconds
kong_kong_latency_ms_count count Count for Kong's own latency
kong_kong_latency_ms_sum ms Sum of Kong's own latency, unit in milliseconds
kong_latency_ms_bucket ms Histogram bucket for request latency, unit in milliseconds
kong_latency_ms_count count Count for request latency
kong_latency_ms_sum ms Sum of request latency, unit in milliseconds
kong_memory_lua_shared_dict_bytes bytes Memory used by Lua shared dictionary, unit in bytes
kong_memory_lua_shared_dict_total_bytes bytes Total memory of Lua shared dictionary, unit in bytes
kong_memory_workers_lua_vms_bytes bytes Memory used by Lua virtual machines in worker processes, unit in bytes
kong_nginx_connections_total count Total number of Nginx connections
kong_nginx_metric_errors_total count Total number of Nginx metric errors
kong_nginx_requests_total count Total number of Nginx requests
kong_node_info - Node information
kong_request_latency_ms_bucket ms Histogram bucket for request latency, unit in milliseconds
kong_request_latency_ms_count count Count for request latency
kong_request_latency_ms_sum ms Sum of request latency, unit in milliseconds
kong_upstream_latency_ms_bucket ms Histogram bucket for upstream latency, unit in milliseconds
kong_upstream_latency_ms_count count Count for upstream latency
kong_upstream_latency_ms_sum ms Sum of upstream latency, unit in milliseconds

Logging

1. Kong Configuration

To enable the reporting of log data, the TCP Log plugin needs to be activated.

Configure via Kong Manager

Log in to Kong Manager, click on the Plugins menu, and add a new TCP Log.

Parameter descriptions:

  • host: DataKit host
  • port: DataKit log socket port

Configure via Kong API

curl -X POST http://localhost:8001/plugins \
    --header "accept: application/json" \
    --header "Content-Type: application/json" \
    --data '
    {
  "name": "tcp-log",
  "config": {
    "port": 9580,
    "host" : "<datakit-host>"
  }
}'

2. DataKit Configuration

  • Go to the conf.d/log directory under the DataKit installation directory, copy logging.conf.sample and name it kong-socket.conf

cp logging.conf.sample kong-socket.conf

  • Modify kong-socket.conf

Replace with the following content:

# {"version": "1.22.0", "desc": "do NOT edit this line"}

[[inputs.logging]]
  ## Required
  ## File names or a pattern to tail.

  # Only two protocols are supported:TCP and UDP.
  sockets = [
     "tcp://0.0.0.0:9580",
  #  "udp://0.0.0.0:9531",
  ]
  ## glob filteer
  ignore = [""]

  ## Your logging source, if it's empty, use 'default'.
  source = "kong_tcp_log"

  ## Add service tag, if it's empty, use $source.
  service = ""

  ## Grok pipeline script name.
  pipeline = ""

  ## optional status:
  ##   "emerg","alert","critical","error","warning","info","debug","OK"
  ignore_status = []

  ## optional encodings:
  ##    "utf-8", "utf-16le", "utf-16be", "gbk", "gb18030" or ""
  character_encoding = ""

  ## The pattern should be a regexp. Note the use of '''this regexp'''.
  ## regexp link: https://golang.org/pkg/regexp/syntax/#hdr-Syntax
  # multiline_match = '''^\S'''

  auto_multiline_detection = true
  auto_multiline_extra_patterns = []

  ## Removes ANSI escape codes from text strings.
  remove_ansi_escape_codes = false

  ## If the data sent failure, will retry forevery.
  blocking_mode = true

  ## If file is inactive, it is ignored.
  ## time units are "ms", "s", "m", "h"
  ignore_dead_log = "1h"

  ## Read file from beginning.
  from_beginning = false

  [inputs.logging.tags]
  # some_tag = "some_value"
  # more_tag = "some_other_value"
  • Restart DataKit

Execute the following command:

datakit service -R

Pipeline

Pipeline can be used for log field extraction. A built-in Pipeline template is available in the Pipeline template library. Find Kong, clone and use it.

Tracing

1. Kong Configuration

Kong needs to enable the OpenTelemetry plugin to report trace data. This feature requires Kong version ≥ 3.3.2.

Enable Global Tracing Configuration

Add the following configuration items to kong.conf: 

tracing_instrumentations = all
tracing_sampling_rate = 1.0

Or add environment variables:

  KONG_TRACING_INSTRUMENTATIONS: all
  KONG_TRACING_SAMPLING_RATE: 1.0

Restart Kong service to take effect.

Configure via Kong Manager

Log in to Kong Manager, click the Plugins menu, and add OpenTelemetry.

Parameter description:

  • traces_endpoint: Reporting address
  • default_format, needs to be consistent with the backend APM protocol, datadog indicates the backend APM uses DDTrace, w3c indicates the backend APM uses OpenTelemetry. Refer to the Kong official documentation for other protocols.

Configure via Kong API

curl -X POST http://localhost:8001/plugins \
    --header "accept: application/json" \
    --header "Content-Type: application/json" \
    --data '
{
  "protocols": [
    "http"
  ],
  "name": "opentelemetry",
  "config": {
    "propagation": {
      "default_format": "datadog"
    },
    "traces_endpoint": "http://192.168.2.110:9529/otel/v1/traces"
  },
  "enabled": true
}'

2. DataKit Configuration

  • Go to the conf.d/opentelemetry directory under the DataKit installation directory, copy opentelemetry.conf.sample and name it opentelemetry.conf

cp opentelemetry.conf.sample opentelemetry.conf

No need to adjust the configuration content

  • Restart DataKit

Execute the following command

datakit service -R

Feedback

Is this page helpful? ×