README.adoc

docker-nodejs-unresponsive-app

Simple Node.js app NOT responding to any TCP request (except for the HTTP maintenance endpoints).

This application is used as a system test helper for circuit breaker scenarios in real environments. Although it is meant as an HTTP application, unresponsitivity can as well be tested with any TCP-based protocol (requires Node v6 or higher).

Startup

Run docker with

docker run -p 8080:8095 docker.pkg.github.com/jforge/docker-nodejs-unresponsive-app/unresponsive-webapp:v-1572443441

The projects matches requirements for a Docker Hub Automated Build, e.g. https://hub.docker.com/r/eolio/docker-nodejs-unresponsive-app/

A corresponding Docker image can be used e.g. for a AWS EC2 Container deployment to get a public available "unresponsive" application very quickly, e.g. http://ec2-35-156-94-40.eu-central-1.compute.amazonaws.com/manage/health

API

Default behaviour is "unresponsive". The server then never responds except for the health checks and responsive switch.

HTTP-GET for the switch is just used for fast testing convenience and does not follow any REST-oriented approach (sorry).

Table 1. Available HTTP Methods

HTTP-Method	Uri	Response	Description
GET	/manage/health	200	Returns health and responsivity status
GET	/api/health	200	Returns health and responsivity status
GET	/manage/set_responsive	200	Sets the behaviour to "responsive"
GET	/manage/set_unresponsive	200	Sets the behaviour to "unresponsive"
GET, POST	/any/other/uri	none (timeout)	Server does not answer with (default) mode "unresponsive"
GET, POST	/any/other/uri	200	Server returns always with 200 with mode "responsive"

"Unresponsive" := the server doesn’t answer at all, if using arbitrary uris except these /api/health and the /manage methods.

Manage Response Payload

The above mentioned /manage methods answer with OK-health and the current (global) responsivity setting.

{
  "status": "OK",
  "responsivity": "unresponsive"
}

Response in "responsive" mode

If the server is set to "responsive", any Uri request but the above mentioned /api/health and /manage methods causes a HTTP 200-OK response together with a payload containing the requested resource.

{
  "status": "OK",
  "requestUri": "/any/other/uri"
}