krakend/krakend-metrics

A set of building blocks for instrumenting KakenD gateways

KrakenD metrics

A set of building blocks for instrumenting KrakenD gateways

Available middlewares

There are the avaliable middlewares to add to the KrakenD pipes.

1. Backend
2. Proxy
3. Router

Available router flavours

1. mux Mux based routers and handlers
2. gin Gin based routers and handlers

Check the examples and the documentation for more details

Configuration

You need to add an ExtraConfig section to the configuration to enable the metrics collector (an empty one will use the defaults).

You can disable metrics by layer by explicitly setting it to true (the default is to be enabled):

• backend_disabled bool
• proxy_disabled bool
• router_disabled bool

Or configure the collection time of metrics:

• collection_time (default: 60s) (Ex: "30s", "5m", "500ms", ...)

Configuration Example

This configuration will set the collection time to 2 minutes and will disable the proxy metrics collector (backend and router metrics will be enabled since the default for all layers is to be enabled).

  "extra_config": {
    "github_com/devopsfaith/krakend-metrics": {
      "collection_time": "2m",
      "proxy_disabled": true,
    }
  }

or leave the defaults:

"extra_config": {
  github_com/devopsfaith/krakend-metrics": {}
}

krakend-metrics's Issues

Suppot disabling a given layer

Right now, the module is enabled by default when the factories are called.

Use the available extra configuration to infer if a layer should be enabled or not.

Support custom configuration

The port and other details regarding the stats recollection should be configurable.

Also, there should be a way to disable the component.

Support complex instrumentation definition

It would be interesting to define a pattern for enabling the instrumentation of a given pipe.

The first filtering strategy that came to my mind is to support regex filtering