このページは、まだ日本語ではご利用いただけません。翻訳中です。

旧バージョンのドキュメントを参照しています。 最新のドキュメントはこちらをご参照ください。

Prometheus metrics

Kong Ingress Controller, as well as Kong Gateway, both expose Prometheus metrics, under certain conditions:

• Kong Ingress Controller, since version 2.0, exposes Prometheus metrics for configuration updates.
• Kong Gateway can expose Prometheus metrics for served requests, if the Prometheus plugin is enabled. See the Using KongPlugin resource guide for information about how to enable a plugin. Also, you can find a specific guide for integration with Prometheus and Grafana as well.

This document is a reference for the former type.

Prometheus metrics for configuration updates

ingress_controller_configuration_push_count

ingress_controller_configuration_push_count (type: counter) provides the number of successful or failed configuration pushes to Kong Gateway.

This metric provides the following labels:

• protocol describes the configuration protocol in use, which can be db-less or deck.
• dataplane describes the data plane that was the target of configuration push.
• success logs the status of configuration updates. If success is false, an unrecoverable error occurred. If success is true, the push succeeded with no errors.
• failure_reason is populated if success="false". It describes the reason for the failure:
  • conflict: A configuration conflict that must be manually fixed.
  • network: A network related issues, such as Kong Gateway is offline.
  • other: Other issues, such as Kong Gateway reporting a non-conflict error.

ingress_controller_translation_count

ingress_controller_translation_count (type: counter) provides the number of translations from the Kubernetes state to the Kong Gateway state.

This metric provides the success label. success logs the status of configuration updates. If success is false, an unrecoverable error occurred. If success is true, the translation succeeded with no errors.

ingress_controller_configuration_push_duration_milliseconds

ingress_controller_configuration_push_duration_milliseconds (type: histogram) is the amount of time, in milliseconds, that it takes to push the configuration to Kong Gateway.

This metric provides the following labels:

• protocol describes the configuration protocol in use, which can be db-less or deck.
• dataplane describes the data plane that was the target of configuration push.
• success logs the status of configuration updates. If success is false, an unrecoverable error occurred. If success is true, the push succeeded with no errors.

Low-level performance metrics

In addition to the above, Kong Ingress Controller exposes more low-level performance metrics, but these may change from version to version because they are provided by underlying frameworks of Kong Ingress Controller.

A non-exhaustive list of these low-level metrics is described in the following:

• Controller-runtime metrics
• Workqueue metrics